Time division protocol for an ad-hoc, peer-to-peer radio network having coordinating channel access to shared parallel data channels with separate reservation channel

[0001] Priority of provision application serial No. 60/246,833, filed on Nov. 8, 2000 is herewith claimed.

[0002] The present invention is directed to a novel protocol for an ad-hoc, peer-to-peer radio network system having coordinating channel access to shared parallel data channels via a separate reservation channel. This system is disclosed in copending application Ser. No. 09/705,588, filed on Nov. 3, 2001, entitled Methods and Apparatus for Coordinating Channel Access to Shared Parallel Data Channels, which application is incorporated by reference herein in its entirety.

[0003] The network system having coordinating channel access to shared parallel data channels via a separate reservation channel of copending application Ser. No. 09/705,588 is directed to a network system, such as radio network, where each node, or radio terminal, of the network is capable of serving as a node or hop of a routing path of a call from another, or to another radio terminal. In that system, communication between nodes or radio terminals is achieved using Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) protocol with the addition of multiple parallel data channels serviced by one reservation channel. By dedicating a separate reservation channel for the multiple parallel data channels, collision-free access by all of the competing nodes or terminals of the service group of the network is greatly reduced. Communications between terminals or nodes is set up by information exchanged on the separate reservation channel, which information includes all of the call set-up information such as data channel desired to be used for transferring voice, video or data, the desired power level of at least initial transmission, messaging such as Request-to-Send (RTS), Clear-to-Send (CTS), Not-Clear-to-Send (NCLS), Acknowledgment (ACK) for indicating reception of the transmitted call, Non-Acknowledgment (NACK) for indicating improper reception of the call, etc. In this system, in order to further ensure fast, adequate and collision-free transmission and reception, besides a primary modem typically provided with the transceiver of each node or terminal, a secondary modem is also provided which is dedicated to the reservation channel when the primary modem of the transceiver is occupied, such as when sending out data on a data channel. This system also provides for collision free transmission and reception between nodes or terminals by transmitting the reservation and data channels in time slots of time frames, with the information as to which time slot is to be used being included in the messaging transmitted by the reservation channel. Such a format not only provides collision-free transmission, but also allows for Quality-of-Service (QoS) for different types of Class-of-Service (CoS), Thus, not only may voice and video be transmitted, besides data, but voice and data transmission may be prioritized, so that when competing calls vie for a data channel, the delay-dependent voice or video transmissions will take precedence. This prioritization is accomplished by assigning prioritized calls for transmission in earlier time slots of a time frame.

[0004] The network system disclosed in U.S. application Ser. No. 09/705,588 ensures that every node or terminal of a service set of terminals has the most information regarding all of other terminals of that service set, so that the choice of data channel to be used, any required delay is transmitting the call, information on power level, and the like, are checked and updated by each terminal by a practically continuous monitoring of the reservation channel.

[0005] As explained above, the system disclosed in U.S. application Ser. No. 09/705,588 utilizes protocol that provides collision-free channel access, which also emphasizes improving geographic reuse of the frequency spectrum.

[0006] In U.S. Pat. No. 5,943,322Mayer, et al., which patent is incorporated by reference herein, the radio system thereof is for use in battlefield conditions. The ad-hoc, peer-to-peer radio system of this patent does not have, nor require, a base station, as conventional cellular systems, personal communications system (PCC), and the like, require; instead, each radio terminal forming part of the ad-hoc, peer-to-peer radio system may alternatively serve as a base station, in addition to being an ordinary link terminal of the radio system, whereby, if one such terminal serving as a base station should for some reason become inoperative, another terminal may take over and serve as the base station. In this patent, personal voice communications is based on a time division duplex (TDD) technique in a code division multiple access (CDMA) system, is operated without a fixed base station, and is provided with simultaneous transmission of a communications channel and a control channel, each spread by different PN codes. The PN code facilitates restricting communications on the network to a particular voice-conversation mode and between identified radios. Transmissions are performed in a time division duplex manner in 62.5 milliseconds slots. One of the radios initiates transmission and maintains power control and time synchronization normally done by a base station. A network control station can voluntarily or by command transfer control of the network to any of the other radios on the network. Colliding transmissions from more than one radio require the radios to retry transmitting until one of the radios transmits in an earlier time slot. Conversational mode capability is provided by equipping the radio receivers with despreaders in parallel for permitting a receiving radio to separately despread the simultaneously transmitted signals all other radios on the network and responding to each radio transmission individually. Simultaneous voice and data communications can be accomplished by equipping the receivers with despreaders for discriminating voice and data information signals spread by different PN codes.

[0007] In commonly-owned provisional application Ser. No. 60/248,182, which application is incorporated by reference herein, there is disclosed an ad-hoc, peer-to-peer radio system for use as a stand-alone system that is also connected to a cellular network and/or PSTN. The ad-hoc mobile radio networking system thereof is capable of receiving and transmitting voice, data and video calls through any number of different types of telecommunication networks, such as the PSTN, the Internet, and the like, besides the cellular and next-generation cellular networks.

[0008] Past research has shown that conventional Carrier Sense Multiple Access (CSMA) algorithms experience diminishing returns when networks approach their ultimate capacity. The vast majority of current research centers on channel access algorithms that provide transmission capacity over a single shared medium. An example of this is the IEEE 802.11 wireless standard which employs a Carrier Sense Multiple Access/Collision Avoidance (CSMA/CA) algorithm. All users within a Basic Service Set (BSS) share a common channel resource.

[0009] The ad-hoc, peer-to-peer radio system of the present invention is based on a transport-mechanism using a time division duplex (TDD) technique in a code division multiple access (CDMA) system. Time Division Duplex (TDD) is a way of maximizing the bits/hz/m2. Such a system not only may be used for providing commercial voice, but is also quite suited to both transmission and reception of data and video services. Time Division Duplex (TDD) systems are typically used for packet data systems, since they make much more efficient use of the available bandwidth, in order to deliver a much higher effective data rate to the end user. TDD is typically used in fixed wired solutions or point-to-point wireless systems because it has its own spectrum limitations. TDD systems, however, have not hitherto been deployed for voice systems.

[0010] Unlike the personal communication radio system of U.S. Pat. No. 5,943,322Mayer, et al., the Time-Division Protocol (TDP) of the present invention does not care about the modem-type of access to radio spectrum, and is designed to work with or without a base station or gateway, since modem functionality is not part of the TDP of the present invention. The protocol of the present invention uses one control/configuration channel and three or more data channels, where communication between radio terminals is planned for preventing interference. Time synchronization is independent of the communication,, whereby no collisions among terminals are possible for configuration data, excepting in the last time slot, and no collisions are possible in the data channels, as described above. The protocol of the present invention may transmit data and video, in addition to voice, since each is just another class of data.

[0011] The system of the present invention is much more complex due to multiple, parallel data channels that are coordinated by a single reservation channel. In this system, a combination of CSMA/CA, TDMA (time division multiple access), FDMA (frequency division multiple access), and CDMA (code division multiple access) is used within the channel access algorithm. The transceiver used in the system employs two modems to solve the channel reliability issues with multiple channel designs, as disclosed in the above-described copending U.S. application Ser. No. 09/705,588. Specifically, the system dedicates a receive-only modem for gathering channel usage information on the reservation channel. The reservation channel operates a hybrid CSMA/CA and TDMA algorithm. The remainder of the protocol uses FDMA for the multiple data channels, and CDMA for multiple users on the same data channel.

[0012] Reference is also had to copending, commonly-owned U.S. patent application Ser. No. 09/______, filed on ______ ______, 2001, entitled Prioritized-Routing for an Ad-Hoc, Peer-to-Peer, Mobile Radio Access System, which is incorporated by reference herein, in which there is disclosed an example of routing table messaging which may be used in the present invention.

[0013] It is the primary objective of the present invention to provide an ad-hoc radio system as part of an overall, larger cellular network, and/or as a stand-alone, independent system, in order to provide commercial use for providing voice, data and video communications between radio terminals of the radio system of the invention and between equipment outside the system of the invention.

[0014] It is also a primary objective of the present invention to provide an overall protocol for ad-hoc radio system not utilizing a fixed base station, whereby a connection path by which a call is made takes into consideration the power loss associated therewith, in order to determine the least-energy routing of a call for the particular service type being transmitted, such as voice, data or video.

[0015] The protocol of the present invention is based on a time-division duplex (TDD) plus code-division multiple access (CDMA) burst packet technology used within the channel access algorithm of the system of the present invention. This provides the improvements in throughput and reliability that are required to deliver high quality voice, video and data. The reservation channel implements a time division multiple access algorithm with dynamic slot allocation. In a distributed manner, nodes determine geographic reuse of slots based on channel quality. Signal quality calculations are used to determine the likelihood of a slot reuse causing destructive interference within a node's neighborhood. Requests for slot usage are compared with the known traffic pattern and accepted or rejected by nodes within RF signal range based on the signal quality calculations. Additionally, the algorithm of the present invention readily provides for the mobility of nodes between geographic areas through the use of a special slot that is reserved for nodes without reservations. Nomadic nodes use this slot to locate a permanent slot to claim for their use. Once claimed, the collision free properties can be enforced to improve the reliability and throughput of messages generated by this node. This results in a maximal use of the spectrum within a geographic area.

[0016] The system of the present invention utilizes a method and algorithm which, in the preferred embodiment, i intended for an ad-hoc network system called ArachNet, and is based on least-energy routing of calls from between network radio terminals. In simple terms, the major component of the routing decision is to choose t route to the destination that uses the least amount of energy over the complete route. The major reason for thi that least-energy routing minimizes the radiated RF energy, in order to reduce interference between terminal A consequence of this is that it creates the most efficient use of the power supply of the terminals. Routing tables based on this least energy routing a developed by the system of the invention, and stored at one or more radio terminals, which routing tables are transmitted and stored by other terminals forming part of the link by which a call is connected. An example of such a routing table is disclosed in copending, commonly-owned U.S. patent application Ser. No. 09/______, filed on ______ ______, 2001, entitled Prioritized-Routing for a Ad-Hoc, Peer-to-Peer, Mobile Radio Access System, which is incorporated by reference herein.

[0017] Variants or equivalents of the system of the invention are possible. There are a number of variants of this approach that would provide acceptable performance. These variants include tuning of each of the four access schemesCSMA/CA, TDMA, FDMA, and CDMA. For example, the width of the time slots may be adjusted based on the specific network over which the protocol is executing. Performance of the network is very dependent on the number of parallel data channels which can be used. A balance exists between the capacity of the reservation channel to make data reservations and the capacity of the data channels to provide service. This balance is dependent on the underlying capabilities of the dedicated, reservations-channel modem that implements the protocol. The performance of the protocol is also dependent on the inclusion of the channel quality extracted from the channel. Accurate estimates of the signal strength translate into improvements in geographic reuse, which can be obtained by aggressive power control schemes. Another example is the use of advancements in the codes used within the CDMA portion. Codes which improve the cross-correlation performance of terminals which share a common data channel improve the throughput and reliability of the overall network performance.

[0018] The adaptive power algorithm of system of the present invention leads to improvements in the determination of RF radius for a given data rate. Increasing the data rate and reducing power promotes geographic reuse. Any loss in communication is easily compensated by our ad-hoc routing algorithms.

[0019] The channel access approach of the invention is equally applicable for subnets which include or do not include gateways. In the gateway approach, time is coordinated within the ad hoc environment by the gateway. In the non-gateway approach, a distributed time algorithm provides acceptable performance. In general, gateways permit the creation of larger networks such as MAN's and WAN's.

[0020] While the protocol method of the present invention is disclosed with regard to an ad-hoc, peer-to-peer radio system, the protocol is equally applicable to any wireless LAN, wireline network, and the like, to which the method and system disclosed in copending U.S. application Ser. No. 09/705,588 may apply.

[0021] The present invention will be more readily understood with reference to the accompanying drawings, wherein:

[0022] FIG. 1 is a logical flow chart showing the software structure of the protocol of the system of the present invention;

[0023] FIG. 2 is depiction of the time-division of the TDD protocol (AP) of the present invention showing the time frames thereof with separated time slots;

[0024] FIG. 3 is a depiction of the many terminals (AT) connected to a particular gateway of the system of the present invention connecting the terminals to an exterior network, with one particular AT entering the service domain thereof and the connection path thereof;

[0025] FIG. 4 is a depiction similar to FIG. 3, but showing the particular AT moving away from the gateway and other AT's, whereby disconnection or reconnection must be carried out;

[0026] FIG. 5 is a depiction similar to FIGS. 3 and 4, showing a gateway connecting the system to outside networks, which connection path does not require very high speed, and where the total energy for passing data along the route is minimized;

[0027] FIG. 6 is a depiction of the time frames (TF's) with time slots (TS's) of the protocol of the system of the present invention required for performing one hop between terminals (AT's) for a permanent link, in order to assure proper communication;

[0028] FIG. 7 is a depiction similar to FIGS. 3-5, but showing the connecting route between two AT's of a local link, in order to control power requirements of the AT's;

[0029] FIG. 8 is a graphical depiction showing an AT approaching a group of AT's and the closed triangular connection therebetween;

[0030] FIG. 9 is a graphical depiction similar to FIG. 8 showing the open triangular connection therebetween if the AT's experience power reduction, whereby connection is preserved via the intermediate AT with ensuing reduction in energy loss;

[0031] FIG. 10 is a graphical depiction similar to FIGS. 8 and 9 showing the perturbed power profile of the AT's and the connection therebetween after the lowering of the power of all of the AT's;

[0032] FIG. 11 is a graphical depiction similar to FIG. 10 and showing the closed triangular connection therebetween after one time frame after power perturbation; and

[0033] FIG. 12 is a graphical depiction similar to FIG. 11 and showing the open triangular connection therebetween after two time frame after power perturbation with the ensuing steadystate, energy-saving path-connection between the AT's.

[0034] For purposes of a better understanding of the description, the following definitions and abbreviations are hereby given:

[0035] Service Area of a Terminal

[0036] The geographical area where the transmission of a terminal can be received at a level higher than environment noise.

[0037] Receive Set of a Terminal

[0038] The set of terminals located within the service area.

[0039] Transmit Set of a Terminal

[0040] The set of terminals containing one particular terminal within their service areas.

[0041] Service Set of a Terminal

[0042] The set of terminals that can receive the transmission of one particular terminal and can be received at that terminal (the intersection between the receive set and the transmit set).

[0043] Simple Connection

[0044] An abstract notion associated to two terminals that can communicate one with another.

[0045] Connecting Path

[0046] A set of adjacent Simple Connections.

[0047] Service Group of a Terminal

[0048] The largest set of terminals containing at least one Connecting Path between the host terminal and any other terminal of the set.

[0049] Set of Active Time Slots

[0050] All time slots used by the service set of a terminal.

[0051] Source Terminal

[0052] The terminal requesting the service.

[0053] Destination Terminal

[0054] The terminal requested to provide the service.

[0055] Route

[0056] The Connecting Path between the Source and the Destination of a service (voice, Internet access or data transfer).

[0057] Link

[0058] The Route, the Service and the transmitting plan at each hop along the route.

[0059] Isolated Network

[0060] A network of terminals not connected to a gateway.

ABBREVIATIONS
AD	Application data
	Data required or generated by an application using AT for data
	transfer. Examples of such applications are: Internet browser,
	telephone, file transfer server/client, Internet games, e-mail send/
	receive, short message services, Internet radio/TV broadcaster/
	receiver; emergency video/audio/text message broadcaster/receiver,
	report of appliance (including automotive) functionality status,
	teleconferencing video/audio participant, etc.
AP	Arachnet Protocol
	The protocol supporting the connection and data transfer between
	AT's.
AT	Arachnet Terminal
	The wireless terminal of the system of the present invention.
ATS	Active Time Slot Set
	The set of time slots that are used by an AT or its service set.
CD	Configuration Data
	Data exchanged in Configuration Channel for maintaining the
	connectivity between AT's.
CC	Configuration Channel
	The radio channel selected for exchanging Configuration Data
	(channel F0).
DC	Data Channel
	Radio channels used for exchanging Application data (channels
	F1, F2 and F3).
GW	Gateway
	The special type of fixed AT that provides connection to the
	World through land (not wireless) connections.
IFTG	Inter Frame Time Gap
	The time gap between the end of the last TS of a TF and the
	beginning of the next TF.
IN	Isolated Network
	The network not connected to the world. An IN has a root AT that
	provides the functionality needed for data routing and connectivity.
LLC	Logical Link Control
	The higher level of the protocol stack providing the interface
	between the network and applications.
MAC	Medium Access Control
	The medium level of the protocol stack providing the control of the
	access to the radio spectrum.
PAL	Physical Access Layer
	The lower level of the protocol stack responsible for transmitting
	and receiving data to/from other AT's.
RxS	Receive Set
	The set of terminals that can receive the signal transmitted by the
	AT owning the set.
SA	Service Area
	The area where the signal transmitted by an AT can be received at
	a level higher than environment noise.
SG	Service Group of an AT
	The group of AT's that can be connected to the host AT with at
	least one connecting path.
SS	Service Set
	The set of AT's that can receive the transmission from the host AT
	and can be received at the host AT.
TF	Time Frame
	A division of the time. The size of TF if configurable and depends
	on several environmental factors.
TS	Time Slot
	A division of the TF. The size and number of TS within a TF is
	configurable.
TxS	Transmit Set of an AT
	The set of AT's that can be received by the host AT.

[0061] The protocol (AP) of the system of the present invention applies to an ad-hoc, peer-to-peer radio network system having coordinating channel access to shared parallel data channels via a separate reservation channel, as disclosed in copending U.S. application Ser. No. 09/705,588. In the radio network system of the invention, there is no fixed base station; each radio terminal is capable of acting as a mobile base station. The protocol of the present invention provides such an ad-hoc, peer-to-peer radio system with the capability of preventing collisions of data transfer. In high-density populated area (conference halls, stadium, downtown of big cities, etc.), the protocol of the present invention allows each terminal to perform close to its maximum theoretical capacity, while dropping the requests in excess. Such behavior is in contrast with conventional polling-type protocols that cannot provide any service when the number of requested connections is larger than a particular fraction of terminal capacity.

[0062] For implementing the protocol of the present invention, each terminal (AT) has fill information about all activities of other terminals and will provide all other terminals with fill information about its own activity. According to the present invention, a combination of TDMA (time division multiple access), FDMA (frequency division multiple access), and CDMA (code division multiple access) is used within the channel access algorithm of the system of the present invention. This provides the improvements in throughput and reliability that are required to deliver high quality voice, video or data. The reservation channel implements a time division multiple access (TDMA) algorithm with dynamic slot allocation. In a distributed manner, nodes determine geographic reuse of slots based on channel quality extracted from messaging in a separate reservation channel. Signal quality calculations are used to determine the likelihood of a slot reuse causing destructive interference within a node's neighborhood. Requests for slot usage are compared with the known traffic patterns, and accepted or rejected by nodes within RF signal range based on the signal quality calculations. Additionally, the algorithm of the present invention readily provides for the mobility of nodes between geographic areas through the use of a special slot that is reserved for nodes without reservations. Nomadic nodes use this slot to locate a permanent slot to claim for their use. Once claimed, the collision free properties can be enforced to improve the reliability and throughput of messages generated by this node. This results in a maximal use of the spectrum within a geographic area.

[0063] The system of the present invention utilizes a method and algorithm for ad-hoc network system that is based on least-energy routing of calls from and between network radio terminals. In simple terms, the major component of the routing decision is to choose the route to the destination that uses the least amount of energy over the complete route. The major reason for this is that least-energy routing minimizes the radiated RF energy, in order to reduce interference between terminals. A consequence of this is that it creates the most efficient use of the power supply of the terminals.

[0064] In a medium dynamically changing its structure, superlative notions as full connectivity, optimal configuration or best structure are, in fact, not applicable, because they cannot be exactly defined. The protocol of the invention makes full use of all available information (that may be incomplete or approximate) about other terminal activities and broadcasts full information about its own current or intended activity. Such cooperative attitude creates the capability to plan and check data-transfer planning before data transfer is initiated.

[0065] In most of the cases, the application data are exchanged between terminals using the same or less transmitting power than the power used for exchanging configuration data on the control or configuration channel. This fact allows better use and reuse of frequencies and time allocation, and makes the application data exchange less sensitive to interference from hidden terminals (AT's).

[0066] For supporting the protocol of the present invention each AT of the radio system has the following capabilities:

[0067] Measures the level of received signal with very good precision;

[0068] Measures the level of received radio noise with very good precision;

[0069] Controls the transmit power;

[0070] Changes fast from receiving to transmitting;

[0071] Changes fast the transmitting or receiving frequency;

[0072] Uses a dedicated receiver for listening to Configuration Channel messages;

[0073] Controls the data rate; and

[0074] Uses a mean for network clock synchronization.

[0075] Software Architecture

[0076] Referring now to FIG. 1, the protocol of the invention is implemented as a three-layer software stack. The lowest, the Physical Access Layer (PAL) 10, is responsible for transmitting and receiving configuration and application data. It exchanges configuration data with the Middle Access layer (MAC) 12 and application data with the highest layer, the Logical Link Control (LLC) 14. Data is received and transmitted according with the communication plans elaborated at MAC. The Medium Access Control (MAC) is responsible for processing received configuration data, for controlling the transmit power, data rate, for creating the data transmit plans and for building the configuration data to be transmitted to the Service Set (SS) of AT terminals. It exchanges routing data with the LLC and configuration data (power, data rate, transmit plans) with the Physical Access Layer. The protocol of the present invention is carried out in this MAC layer. As will explained hereinbelow, when there exists an isolated network (IN) of terminals, the protocol of the invention is capable of being carried out by each respective AT. Control (LLC) layer is responsible for exchanging application data between applications and the Physical Access Layer (PAL). Data received from PAL is unpacked, decrypted and distributed to applications. Data from application is encrypted, packed (adding routing information), and passed to the PAL to be transmitted.

[0077] The exchange of configuration data (CD) between AT's participating in a network of the present invention is accomplished in configuration channel using frequency F0. The other channels (F1, F2, F3) are used for transferring application data (AD) between AT's, thus constituting FDMA scheme of the multiple aspect protocol of the invention.

[0078] The Protocol of the present invention uses a time division scheme (TDD) for organizing the access to airwaves. Referring to FIG. 2, the time is divided in time frames (TF) 16, and each frame is divided in time slots (TS) 18. At the end of each time frame is the Inter Frame Time Gap (IFTG) 20 that has a different length than the regular TS. During the IFTG, no data is sent out by any AT, so that each AT processes data collected from other AT's during the time frame, and will perform required calculations, such as power level, data channel connectivity, etc.

[0079] A terminal (AT) can transmit configuration data in F0 only during its own assigned Time Slot (TS). Initially, an AT signals its presence using the last TS of the time frame. With the next time frame, it must relocate to another TS with a lower rank, that is one earlier in the time frame. This relocation policy reduces substantially the possibility of collisions in the configuration channel. If two or more AT's try to start working during the same time frame (TF), their transmissions may collide, but the collision is identified and corrected by means of conventional PN coding (CDMA). The probability of collision in the data channels (DC's) is almost zero for AT's, since the reservation channel information exchange has already ensured such a collision-free transmission, either by way of the chosen data channel for transmission (FDMA).

[0080] The power level of the modem for the configuration channel (CC) information is greater than that of the modem for transmitting data on the data channels (DC), since an AT must first send out connectivity information with enough power to reach other AT's of its respective service set (SS). Once this has been done, and a routing path determined, which routing path will indicate the first AT that shall constitute the first hop or link of the routing path, which hop is closer to the requesting AT than at least most of all of the other AT's of the SS, the other modem dedicated to the transmission of data on the DC's will only have to transmit at a power level less than that of the modem dedicated to the configuration channel. Thus, since applications data (AD) are transmitted at a lower power than that of the configuration data (CD), the condition for collision in data channels can be identified before it occurs, with appropriate measures being taken for preventing it, such as the use of CDMA. In addition, since the data channel data is transmitted at a lower power level, interference is reduced since the RF waves of the data channels do not propagate as far along the SS. It is noted that in the case where the primary modem is used most of the time for transmitting both configuration data as well as channel data, with the dedicated reservation-channel modem only being used when the primary modem is occupied with sending out messaging on the data channels, the primary modem will have its power level changed in accordance with which channel it is transmitting, as disclosed in copending U.S. application Ser. No. 09/705,588. However, in the preferred form of the invention, the dedicated configuration-channel modem receives and transmits configuration data regardless of the state of the primary modem.

[0081] At very heavy loading, the degradation of the service provided by the protocol of the present invention is expected to remain constant, in contrast with prior-art polling-type protocols that collapse abruptly in similar conditions.

[0082] When first powered on, or when approaching a group, the new terminal (AT) listens to messages in the time frames (TF), creates a utilization map based thereon, and computes its transmit power, in the manner disclosed in copending U.S. application Ser. No. 09/705,588. According to the protocol of the present invention, it submits the first message in the last time slot (TS) of the time frame, using as much power as needed in order to reach all AT's from which it has received similar messaging, that is its service set (SS). The message shows the utilization map it knows about, and requests to register with the closest AT. In the utilization map, it marks as busy all time slots (TS) during which a message or high-level noise was received during the last time frame, and also marks the time slot where it intends to move to with the next frame. The TS where it wants to move in the next time frame will have been reported as free in utilization maps of all AT's of the SS. In every time frame, the AT creates the utilization map based on time slots it identified as being busy (a signal was received during the TS), and it receives similar maps from all other AT's in the transmit-set of each AT (TxS). Identifying free TS's consists in making a bit-wise OR between all received maps. The result shows free time slots as bits with value zero and busy TS as bits with value one.

[0083] The Configuration Channel (CC) is used for passing two kinds of messages: connectivity and data transfer plans. All messages in the connectivity group contain the utilization map, the power used for transmitting the message, and the level of environment noise at a transmission site, beside other, specific, conventional information. These messages register, un-register, and communicate the respective AT status. The status message is transmitted whenever no other message is pending in order to maintain connection.

[0084] The group of messages for data transfer planning is used for adjusting the transmit power, building, re-building, re-routing and releasing links, as described hereinbelow in detail. As disclosed in copending U.S. application Ser. No. 09/705,588, some of them are used before starting the transfer of data packet, and some are used while the data transfer takes place. Data Channels (DC's) are mainly used for moving data packets from one AT to another. Some of the data transfers require confirmation/rejection of received data, and some not. A rejection of received data is an automatic request for retransmitting the associated data package. Broadcast services do not require any confirmation of received data correctness.

[0085] Connected Network

[0086] In order to talk to the world, each AT should be connected directly or indirectly to a gateway that connects the AT's of the service group of AT's (SG) to an outside network, such as a cellular network, PSTN, and the like. When it is connected indirectly, the connectivity is realized through another AT or AT's. An AT loses its connectivity if an uplink AT (an AT closer to the gateway along the connection path) becomes out of range (cannot be heard anymore), or if the uplink AT loses its connectivity. The AT so losing its connection will look for the closest (smallest path loss) connected AT providing the smallest path loss of power, and reconnect through it. If no connected AT is found in the current service set (SS), the disconnected AT will send out status messages every time frame (TF). The transmit power of that AT is increased one dBm every other TF, until another, connected AT answers back. If after reaching the maximum transmitting power (28 dBm) no connected AT can be included in the SS, the AT and its SS are considered as isolated. An SS can be isolated only if the SG containing the SS is isolated, also. An isolated AT will adjust its power according with the power and space topology of its SS. Periodically, isolated AT's will transmit messages using the maximum transmitting power until it is heard by a connected AT that provides the connectivity to the world via a gateway. While an SG is isolated, the services can be provided only between those terminals-members of the service group (SG). An AT that has no service set (SS) will send out messages every 60 seconds using the maximum power. The self-testing functions will be activated before sending out any high-power message to verify hardware viability, since improperly working AT's can disable the network by sending out interfering signals.

[0087] When the AT is powered on, it listens to the transmit set (TxS) from other AT's. It identifies the path loss for each TxS member by subtracting the strength level of the received signal from the transmit power. The highest path loss is used for setting the current transmit power m the configuration channel (CC). The new AT submits a registration request to the closest (smallest path loss) connected AT in the last TS, as described above. The registration request is forwarded to a gateway for use by the LLC layer software. Each terminal along the path remembers the fact that it helped to register the new AT. The first uplink AT is responsible for monitoring the activity of the newly registered AT, and submits a request to unregister it in case it becomes out of range, or if it was heard requesting registration with another AT.

[0088] Referring to FIG. 3, AT7 is shown entering the network. The connecting process and the information held at each AT in this network is as follows:

[0089] AT7

[0090] 1. Submits the registration request to AT6;

[0091] 2. Monitors the evolution of path loss to all received AT's and adjusts the transmit power according with path loss variation and noise level in each AT area; if AT6 is predicted to get out of range in next 5 seconds, AT7 searches for another AT to connect through;

[0092] 3. Monitors the registration status of the AT6.

[0093] Connectivity data: UplinkAT7 Downlinknone:

[0094] AT6

[0095] 1. Submits the received registration request of AT7 to AT3;

[0096] 2. Monitors the path loss to all received AT's and adjusts its own transmit power according with path loss variation and the noise level in each AT area; if AT7 becomes out of range, it submits to AT3 the request to unregister AT7;

[0097] 3. Monitors the activity of AT7; if it identifies that AT7 requests registration with another AT, AT6 submits to AT3 the request to unregister AT7;

[0098] 4. Monitors the evolution of path loss to AT3; if AT3 is predicted to get out of range in next 5 seconds, it searches for another AT to connect through;

[0099] 5. Monitors the registration status of the AT3.

[0100] Connectivity data: UplinkAT3 DownlinkAT7:

[0101] AT3

[0102] 1. Submits the received registration request of AT7 to Gateway;

[0103] 2. Monitors the path loss to all received AT's and adjusts its own transmit power accordingly; if any of AT4, AT5 or AT6 becomes out of range, submits to Gateway the request unregistration the AT;

[0104] 3. Monitors the activity of AT4, AT5 and AT6; if any of them requests registration with another AT, submits the un-register request to the Gateway;

[0105] 4. Monitors the evolution of path loss to Gateway; if it is predicted to get out of range in less than 5 seconds, it searches for another AT (or gateway) to connect through.

[0106] Connectivity data: UplinkGateway DownlinkAT4, AT5, and AT6/AT7:

[0107] Gateway

[0108] 1. Submits registration of AT7 to the global database;

[0109] 2. Monitors the evolution of path loss to all AT's it can receive and adjusts its own transmit power accordingly.

[0110] Connectivity data: Uplinkworld DownlinkAT1/ . . . , AT2/ . . . , and AT3/AT4, AT5, AT6, AT7.

[0111] The registration process creates a tree structure rooted at the gateway. Each AT knows the uplink through which it registered, and the list of direct downlinked AT's from the AT requesting registration. To each direct downlink is associated the list of AT's registered through it. If any AT is turned off or loses the connection with its uplink, the uplink submits the request to unregister the AT that has lost the connection and all its downlink AT's. The unregistration requires all AT's receiving and transmitting to remove the information about the unregistered AT's from their lists.

[0112] The Gateway sends to the Global Database only registration requests. As a result, the Global Database remembers the Gateway where the AT was connected last time.

[0113] Referring now to FIG. 4, it is shown the case where AT6 is moving away from AT3, with which it had been registered and by which it had been connected to the gateway 22. As result, it requests to register with AT5 by sending a message thereto. AT3 listens to this request, and submits to Gateway 22 the request to unregister AT6 and AT7. If AT3 cannot hear AT6 requesting to re-register, it means that the AT6 is out of the range of AT3 and that AT3 must submit to Gateway 22 the request to unregister AT6 when the condition occurs. If the request from AT5 to register AT6 comes before the time-out, AT3 has only to update the connectivity data by moving the AT6 downlink list to AT5 downlink list.

[0114] If AT3 can hear AT6, the process develops as follows:

		Transmit	Transmitter
Frame	Transmit	data	Update	Receiver	Receiver Update
1	AT6	Self	Replace uplink	AT5	Open AT6 downlink
		register	AT3 with AT5		list
		with AT5
1				AT3	Move AT6 downlink
					list (AT6 and AT7)
					to un-register list
2	AT3	Un-register	Remove AT6 from	Gateway	Remove AT6 from
		AT6	un-register list		AT3 downlink list
2	AT5	Register		AT3	Add AT6 to AT5
		AT6			downlink list
2	AT6	Register		AT5	Add AT7 to AT6
		AT7			downlink list
3	AT3	Un-register	Remove AT7 from	Gateway	Remove AT7 from
		AT7	list		AT3 downlink list.
3	AT5	Register		AT3	Add AT7 to AT5
		AT7			downlink list
4	AT3	Register		Gateway	Add AT6 to AT3
		AT6			downlink list
5	AT3	Register		Gateway	Add AT7 to AT3
		AT7			downlink list

[0115] If AT3 cannot hear AT6, the process develops as follows:

	Transmit	Transmit	Transmitter AT	Receiver
Frame	AT	data	Update	AT	Receiver Update
1	AT6	Self	Replace uplink	AT5	Open AT6 downlink
		register	AT3 with AT5		list
		with AT5
2	AT5	Register		AT3	Add AT6 to AT5
		AT6			downlink list. Since
					AT6 was a direct
					downlink, all its
					downlink list is moved
					to un-register list.
2	AT6	Register		AT5	Add AT7 to AT6
		AT7			downlink list
3	AT3	Un-	Remove AT7 from	Gateway	Remove AT7 from
		register	un-register list		AT3 downlink list.
		AT7
3	AT5	Register	Add AT7 to AT6	AT3	Add AT7 to AT5
		AT7	downlink list		downlink list
4	AT3	Register		Gateway	Add AT6 to AT3
		AT6			downlink list
5	AT3	Register		Gateway	Add AT7 to AT3
		AT7			downlink list

[0116] The unregister list contains all AT's to be unregistered. No registration request is processed if the unregister list is not empty.

[0117] All operations described herein are executed in the configuration channel (CC) only. If any of the AT's involved in connectivity update are supporting data transfer at the time, the data transfer is not affected.

[0118] Isolated Network

[0119] When powered on, each terminal is isolated. The AT tries to find a registered AT and register to the world through it through a last time slot of a time frame, as described above. The search consists in sending out status messages and listening for answers. At first, the AT listens in the configuration channel (CC). If no other AT can be heard, it transmits the status message in a time slot (TS) randomly selected. With every time frame (TF), it increases the transmitting power until a response is received. The response may be from a gateway 22 or other AT's. In the next step, the requesting AT requests the registration with one of the correspondents in the following priorities: The closest gateway, the closest AT registered with the world, or the closest AT registered with an isolated network. If no response is heard and the transmitting power is at maximum level (28 dBm), the AT is, therefore, isolated, and becomes the root of an isolated network (IN). The identification of the IN is the time in seconds since the Jan. 1, 2001 when the AT was powered on, or another similar method.

[0120] All members of an isolated network (IN) send out the status message at maximum power level (28 dBm) at a random rate, varying between 10 and N seconds, where N is three times the number of members in the IN. The message is sent in the last time slot (TS) of the first time frame (TF). AT's in the same IN do not adjust their power based on the message received during the last TS of the first TF. The message is intended to identify the possible connectivity to another IN or to connected networks. The AT, including the root, that can hear AT's from more than one isolated network (IN), should request registration with the closest AT member of the IN with the largest identification number (the older). This method will create in an IN a tree structure similar with the structure of a connected network (CN)

[0121] Routing

[0122] Most of the time, AT's are connected to the world through a gateway or gateways 22. In some particular cases, a group of AT's can be isolated from the world, if no functioning gateway is available. Routing data through an isolated network (IN) is no different than routing data through a connected network (CN).

[0123] Powering a Link

[0124] For supporting the transfer of data between AT's, the protocol (AP) of the present invention uses the concept of a link. The link is the selected route connecting the source AT to the destination AT or gateway, and includes: The type of service provided, such as voice, data or video; the time slot (TS) used on each AT hop for transporting data; and, indirectly, the application instantiation. Between an AT and a gateway there may be active in the same time many links each supporting another application or instance of the same application. Application data is usually transferred between the AT and associated gateway by the link, using a sequence of many AT's. The AT-connectivity process creates a connectivity path between the terminal and its gateway. The connectivity path uses the smallest possible power, which, therefore, implies the use of a large number of hops and a large pipeline delay, which is permissible when the class of service is data transfer.

[0125] Referring now to FIG. 5, there is shown the connection path of: GatewayAT3AT5AT6AT7. This may be used to support the link that passes data for applications not requiring very high speed, or accepting large delay or latency. Thus, the total energy used for passing data along this route is the smallest possible. Over the connectivity created by the longest path, one has the route GatewayAT3AT7 that has only two hops, and may be used for exchanging data with applications requiring smaller delays such as voice or video transmission. The connectivity route GatewayAT7 has one hop only, but it requires much more energy than any other route. A high-energy route implies the use of high transmit power for transmitting the data. According to the present invention, in order to prevent unexpected interference, the transmitting power in the data channels (DC's) are the same or lower than the transmitting power in configuration channel (CC). In order to ensure the full cooperation between AT's, the whole service group (SG) will adjust its power in the configuration channel (CC), even if only one route has a real need for it, as described hereinbelow.

[0126] The use of high transmit-power has two side effects. It drains the battery of mobile AT's faster, and reduces the availability of system resources, making it difficult to reuse frequency and time slots. If the connection path between the gateway 22 and the client AT has N₁ hops, and the gateway power is P₁, and the length of the connection route should be no more than N₂ hops, the new power to be used at each end of the path is P₂ $\begin{array}{cc}{P}_2=P_1+30{\log }_{10}\left(\frac{N_1}{N_2}\right)dBm& \text{(0-1)}\end{array}$

[0127] Equation (0-1) provides a means to compute the new, greater power P₂ that should make the path to have only N₂ hops. The parameter is the space absorption factor. Its value is dependent on many factors, including the propagation media characteristics, such as free space, concrete walls and floors, wooden walls, metal frame structure, foliage, and the like, lateral reflections, vertical reflections, etc. The initial value for may be 1.0, but it should be adjusted based on system reaction to the intent to the changing of the number of hops.

[0128] The corrected power is applied at the gateway and at the client AT, a fact that attracts automatic change of the power profile along the entire connection route. If the correction does not have the expected result, a second correction will be applied after the route has been established.

[0129] Messaging Based on Least Energy Routing

[0130] The protocol of the present invention is based on least energy routing determination, as discussed previous especially when transmitting data. The routing table messaging that is exchanged between termimals may ha format as that disclosed in copending, commonly-owned U.S. patent application Ser. No. 09/______ , filed on ______ ______, 2001, entitled Prioritized-Routing for an Ad-Hoc, Peer-to-Peer, Mobile Radio Access System which is incorporated by reference herein.

[0131] of the protocol of the invention is used to set up the optimal path of a call. The following algorithm of the protocol of the present invention is based on this minimum energy routing.

source-routing (message_ptr,msg-length,destination, msg-type)
/* source based routing including link adaption algorithm
*/
opt_route(destination, msg_type)
/* determine optimal route to destination this will return the best available route
based on Class Of Service (COS) from msg_type and other network parameters
including link quality. The returned information will be used to calculate the data rate
and power level
*/
calc_symbol_rate (sym_rate)
calc_code_rate (code_rate)
calc_pwr_level (pwr_level)
send_msg(RTS,msg_length,destination,sym_rate,code_rate,pwr_level)
/* send RTS to first router and await CTS to send the data packet
opt_route (destination, msg_type)
RTS refers to Request-To-Send message; CTS refers to Clear-To-Send message; msg
refers to the message sent from each terminal.
/*The following algorithm determines the best route to the destination based on the COS
in the message type.
The following example illustrates the decision process:
Route1 term1-> term4
Low latency, BER high
Route 2 term1 -> term2-> term4
High latency, BER low
Route 3 term 1-> term2-> term3-> term4
High latency, BER low
Route 4 term 1 -> term5 -> term6 -> term4
Low latency, BER low
BER is Bit-Error-Rate; latency is delay.

[0132] In the case of a voice call that has a COS that can tolerate a high BER but not high latency, it would choose route 1 over route 4 because it cannot tolerate high latency.

[0133] In the case of a data call that has a COS that can tolerate high latency but not high BER, it will choose route 2 or route 3.

[0134] Linking

[0135] The protocol (AP) of the present invention defines permanent and temporary links. A permanent link remains active until it is changed or released, while a temporary link is used only once. Permanent links are used for transmitting any type of information, such as voice, data and video. The subsystem providing the information may not be able to provide it at a constant rate while the link is being planned. If there is no information to be transmitted when the transmission time comes, the AT sends a maintenance package which has no other reason other than to maintain the link active, and to give the next hop the opportunity to measure and reply regarding transmission quality. The maintenance packages are also created at a gateway 22 when the land-network is late and provides no data, or at any AT that receives incorrect data and has no data in link queue. The maintenance packet is dropped if the receiving AT has data packets in link queue. Otherwise, the maintenance packet is sent to next AT in the link. A hop of the link is released (removed from AT data structures) when no data, no maintenance packet and no noise is received in the first time slot (TS) of the link hop. The lack of transmitted data is always followed by an abundance of data exceeding the capacity of the planned link. Packets of data have to be stored and planned for transmission in proper sequence order. A configuration parameter, combined with measurements of delivery rate, is used for identifying when the amount of accumulated data requires special action. Such action could be to drop some data in excess, or to send out some of it using temporary links.

[0136] The permanent link is initiated when a service calls the client AT or the client AT calls a service. The source AT and the gateway identify the depth, the number of hops between the gateway and the AT, of the client AT. The required number of hops is computed according with service requirements. The gateway and client AT transmitting power in the configuration channel (CC) is computed using equation (0-1). For each hop, the power in the data channels (DC) is computed based on power loss to the next hop. If needed, the power used in the configuration channel (CC) is increased, such that it is at least 2 dBm higher than the data-channel transmits power. The transmit plan is built using only free TS's and TF's. The link message transmitted to the next hop contains the transmitting plan. If any member of the service set (SS) identifies any conflict in the transmitting plan, it answers back showing the channel and the map of the time frame (TF) conflicting with the plan. In such case, a new plan is issued using the newly achieved information. When the plan has no conflicts, the next hop accepts it by issuing a confirmation.

[0137] Referring to FIG. 6, it takes at least two time frames per hop for making the connection. In the first time frame (TF), the transmitting AT sends out the transmit plan as a Request-To-Send (RTS) message. In the next TF, it may receive rejections from other AT's, including the next AT in the route as Not-Clear-To-Send (NCTS) messages. If there are rejections, the transmitting AT issues a new transmit plan and sends it out in a third TF. The receiving AT can send the Clear-To-Send (CTS) message in the third TF, but it also has to listen if no transmit plan was reissued during the same frame. For assuring proper communication, the transmit-plan issued at AT11 must not use for transmitting data the time slots (TS) that AT11 and AT12 are using for communication in the configuration channel (CC). When the permanent link is not needed anymore, a message that requests releasing the link is issued.

[0138] A temporary link is initiated when the amount of informational data pertaining to a permanent link passes over a predefined limit, and it is too large to cover the network jitter. The initiator AT sends a request for a temporary link which includes the transmit plan. If any member of the service set (SS) finds any conflict in the plan with the currently assigned time frame (TF) and time slot (TS), the information on conflicting data channel is transmitted. The issuer AT has to make another plan and re-transmit it. The next hop confirms the plan. It takes at least three TF's to fully define a temporary link. It means that the temporary link must target time slots that are in time frames at least three TF's ahead.

[0139] Local Link

[0140] The local link is defined as a link between two AT's connected to the world through the same gateway or two members of the same isolated network (IN). FIG. 7 shows the connecting route between AT4 and AT7. AT4 requested the link to AT7. Depending upon the type of required service, it increases its power, and sends the request for power adjustment to AT7. The request is used for identifying a connection route between the source and destination, and to control the power at both ends based on service requirements. Since AT4 does not know where AT7 is located, it issues the power control message toward the gateway or the root AT. AT3 receives this request and finds AT7 in the AT5 downlink list. It then directs the message toward AT5. AT5 directs the request to AT6, and then AT6 directs it to AT7.

[0141] In local links, the request is always started at an AT, which sends it toward the gateway or root AT. While the requests advance, the destination AT is checked in local lists of each node. If it is found, the link-request is routed toward the destination. If the link-request arrives at the gateway and the destination AT is not in gateway list of registered AT's, the request is passed to the world that may reject it or connect to the service provider. In contrast with the gateway, the root of isolated network (IN) rejects the link if the destination AT is not in its registration list.

[0142] Building the Link

[0143] The process of building a link has two steps. In the first step, the request for setting the link power travels from the source to the destination. The trace of the message is saved in each AT along the connection route. The message carries information about the length of the path, which is incremented while the message is passed from one AT to the next. The power control request does not require confirmation. The sender listens in the next time frame (TF) if the next hop retransmits the message. If it does not, the message is repeated. The source increases its transmit power when sending out the power control request based on its distance to the gateway, or the root AT if in an IN. When the message arrives at the end of the connection path, the destination increases its transmit power in accordance with length of the path and service requirements. Then the destination sends out a dummy Clear-to-Send (CTS) message using the new power. As response to this message, all AT's that can hear it, and were part of the connection path for the link, answer with their Ready-To-Link (RTL) message. If the destination made a substantial change in transmitting power, it has to wait several frames until the service set (SS) stabilizes. After that, the source AT can select the proper AT from all of the answers from AT's in order to build the first hop, based on that AT's position along the connectivity route. While the link-hop is created using RTS/CTS/NCTS messaging, all AT's along the connectivity path that can receive the RTS/CTS/NCTS messages answer with a RTL message. When the second hop is ready to be built, the second AT in the link has data for selecting the next AT in the path.

[0144] In FIG. 7, it is shown that AT6 answered the dummy CTS submitted by AT7 while it was increasing its power. Then AT7 issued the RTS to AT6. While AT6 was confirming the link-hop by sending a CTS message, AT4 and AT5 received it, and answered with a RTL message. AT6 selected AT4 for the next hop because it has the smallest distance from the source as opposed to AT5.

[0145] Building the link takes a relatively long time in the protocol (AP) of the present invention, but after it is built, the link remains active for a long time. Disconnections due to network mobility are repaired while information is transferred, and do not require the redoing of the whole procedure. The power of each AT participating in the respective link is adjusted dynamically in order to maintain the proper service-quality and network connectivity.

[0146] Rerouting

[0147] For AT's not supporting a link, the transmitting power in the configuration channel (CC) is computed based on the AT's relative positions within the service group (SG). As an AT moves, it may re-register with the same or with another gateway. An active link is more rigid and follows the changes in the connectivity tree at a slower pace. The AT's supporting the link have their power in the configuration channel (CC) and data channels (DC) controlled by the group topology and the type of supported service. When an AT predicts that its uplink- connection will go out of range, or it is already out of range, it changes its uplink-connection and registers with another AT. The new registration may or may not change the gateway to which the AT was previously registered. After the registration request, the AT sends out the request for setting the transmit power and the request for rerouting the link(s). Then, it registers all sources and destination of supported links. Finally, the AT sends out the request to register all other AT's that have registered through it. Information from the gateway starts flowing through the new route as soon as it is connected. Information towards the gateway is directed to the new route after a delay of time frames depending on the type of service provided. For slow connections, such as data, the delay is twice the requester distance to the gateway. For fast connections, such as voice and video, the delay is equal with the requester distance to the gateway.

[0148] Re-link

[0149] Each data transmission is confirmed with an acknowledge (ACK) message. NACK message is used for marking improperly received data that have to be retransmitted. The failure of data reception can be caused by multi-path, co-channel or adjacent channel interference. Usually, lowering the transmission rate solves the multi-path interference. Re-planning the link segment solves the co-channel and adjacent channel interference. A RELINK message is used for establishing a new transmit plan between two adjacent AT's of the same link.

[0150] Length Adjustment

[0151] The destination AT of the link can measure the delay and can decide to reduce it. The adjustment process starts with increasing the power in the configuration channel (CC) and sending the power-request message to the source. A dummy clear-to-send (CTS) message is issued for selecting the AT for the first hop of the link. The re-link request is sent to this AT that continues the process, until it reaches the source. The re-link request travels against the informational flow. Information is directed onto the new segments of the route at the AT's shared by the new and old routes. The old segments of the route are released, since no data is sent through them anymore.

[0152] Regarding data packets, since they have sequence numbers, after rerouting a link, the destination AT could receive the data packets in incorrect sequence. It must send them in proper sequence. Therefore, late, out-of-sequence packets get the highest transmission priority.

[0153] Power Control

[0154] The control of power is important for maintaining the connectivity between service group (SG) members and the quality of informational transfer.

[0155] Oscillation

[0156] The service group (SG) of AT's using the protocol (AP) of the present invention has the tendency to stabilize at a power profile that reflects the relative path loss between terminals. This stabilization is not 100% accurate, because decisions made at one time frame (TF) are based on measurements made during the previous TF. At the time the decision is applied, the group has already changed its status. For this reason, the group power-profile may have oscillations around its stable position. For preventing this oscillation, the transmit power in the configuration channel (CC) is filtered using the power used in last three time frames (TF's). If there is a repeat of the transmit power-level in last three TF's, the current transmit power-level is computed with the average between the currently computed power-level and the transmit power-level used in previous TF; otherwise the last computed power-level is used as transmit power-level.

[0157] Variation

[0158] In the configuration channel (CC), each AT listens to all terminals and measures the level of the received signals. The difference between the transmit power-level that is part of the received message and the power-level of the received signal provides the measure of the loss of signal due to propagation. After listening to its service set (SS), a AT selects the largest path loss and adjusts its power-level to be able to reach that terminal at a power-level higher than the local noise. The protocol of the-present invention makes the supposition that the path loss between any two terminals is symmetrical, excepting for local noise. In reality, the computed signal loss is not symmetrical in both directions, as receiver sensitivity and transmitter efficiency have variations from one terminal to another due to parts variety, manufacturing process, tuning, or terminal aging. The protocol (AP) of the present invention can properly control channel access if these variations are less than 5 dB

[0159] Each AT keeps historical data about path-loss evolution for predicting the connection-status in next 3-5 seconds. Equation (0-2) provides a simple method for acquiring the path loss variation, as described below. From measurements P(t₁) and P(t₂) of the path loss at time t₁ and t₂, the average path loss variation (t) is computed using an IIF: $\begin{array}{cc}\left(t_2\right)=\left(1-k\right)\left(t_1\right)+kP\left(t_2\right)-\frac{P\left(t_1\right)}{t_2-t_1}& \text{(0-2)}\end{array}$

[0160] The factor k has a very small value (0.01 for example) that are identified empirically. If the measurements are performed every time frame (TF), the difference t₁t₂ is always one. Then the equation becomes:

[0161] In this equation, P and P are the values of path loss measured in the current and the previous time frames (TF's), and and are the values of the average variation of the path loss for the current and previous TF.

[0162] The average value of the path loss is computed with equation (0-4):

[0163] In this equation, L is the average path loss, L is the previous average path loss, P is the last measured path loss, and k is the same filtering constant as before. The predicted value of the path loss after m time frames is computed with equation (0-5):

[0164] If the predicted value of the transmit power that is computed based on the predicted path loss and the noise at correspondent AT is larger than 28 dBm, the connection will be lost in m time frames. Since the increase of power in one AT can cause the whole group to increase the transmit power, the AT may decide to re-register using another AT if it is possible, where such decision may reduce substantially the transmit power. After re-registering, the AT that supports at least a link should request to reroute it.

[0165] Optimization

[0166] When an AT moves away from its service group (SG), the transmit power is increased in order to maintain the connectivity. When an AT moves closer to its SG, the connectivity is preserved, if using the current power-level, but this may not be economical. In FIG. 8, there is shown AT13 approaching its service group (SG). The transmitting power of all members of this group is high enough to allow each AT to communicate with the other two. If AT11 and AT13 reduce their power, they can still be connected through AT12. The condition supporting the decision to lower the power of AT13 and AT11 is that AT12 must be able to talk with both of them.

[0167] In FIG. 9, there is shown the transmit area of each AT after lowering the transmitting power. AT12 is located in the intersection of transmit areas of AT11 and AT13, while AT11 and AT13 are in the transmit area of AT12. It means that at this power profile, AT12 can communicate with AT11 and AT13, but AT11 and AT13 cannot communicate with each other. The system's total transmit power is smaller than it was before. The procedure can be repeated until no triangle is formed. This procedure requires that AT11 and AT13 know that both of them can talk to AT12. The information must be achieved by listening to network talk which can be incomplete at any time.

[0168] The same effect can be achieved through a much more simple procedure. The system of AT's has the tendency to stabilize at a particular power profile. After applying a perturbation, the system returns to the same or to another stable state. If the perturbation consists in lowering the power, the new stable position will be, conditions permitting, at a lower power. As with the triangle method, the perturbation should be applied to all AT's in the system at the same time.

[0169] In FIG. 10, there is shown the situation after lowering the power of all AT's by one dBm. While using the perturbed power profile, AT11 can hear only AT12, AT12 can hear AT11 and AT13, while AT13 cannot hear anybody. Based on this situation, AT11 identifies that it can remain connected to AT12 while using a lower transmit power. AT12 identifies that its transmit power is too low for maintaining the connection to AT13. AT13 considers itself isolated because it cannot hear anybody, and, therefore, increases its transmit power by 1 dBm.

[0170] In FIG. 11, in the next time frame, AT12 and AT13 return to the power level they had before the perturbation, while AT11 uses only as much power as needed for remaining connected to AT12. In the next time frame, AT11 can hear AT12 and AT13, but it uses too little power to transmit to AT13. It decides to increase to the level it was using before the perturbation. The oscillation filter identifies it as a possible oscillation and does not allow an increase larger than half of the increment. AT12 can hear both AT11 and AT13 and finds that its current transmit power is correct for talking with both AT's. AT13 can hear only AT12. It finds out that the power it uses for such connection is too large, and computes the new, lower transmit power using data collected from AT12.

[0171] In FIG. 12, there is shown the situation two time frames after applying the perturbation. AT12 and AT13 are both using the right power level for providing the connection. AT11's transmit power is too high to talk to AT12, the only AT it can hear. With next time frame, AT11 will reduce the power to the proper level. After that reduction, all AT's will have the same transmit power as the power computed using the triangle method.

[0172] Both methods require that the algorithm be executed in the same time frame on all AT's. The triangle method can be applied every time frame or only at predefined times. The perturbation method can be applied only from time to time, but no sooner than 5 time frames, to allow the group to get in a stable position.

[0173] The triangle method provides the final power profile after one time frame, but it requires special computation for identifying which triangle can be broken. The information needed for this method is collected while listening to other AT's talk, a fact that may not be possible when supporting an active link.

[0174] The perturbation method requires three time frames to get to the right power profile. It does not require special computation, as the power control algorithm is run every time frame anyway.

[0175] While a specific embodiment of the invention has been shown and described, it is to be understood that numerous changes and modifications may be made therein without departing from the scope and spirit of the invention as set forth in the appended claims.