WO2007034387A2

WO2007034387A2 - Method for communicating between a plurality of nodes using a wireless communication protocol, inter-vehicle network

Info

Publication number: WO2007034387A2
Application number: PCT/IB2006/053304
Authority: WO
Inventors: Marco Ruffini; Hans-Jürgen Reumerman
Original assignee: Philips Intellectual Property & Standards Gmbh; Koninklijke Philips Electronics N. V.
Priority date: 2005-09-26
Filing date: 2006-09-15
Publication date: 2007-03-29
Also published as: WO2007034387A3

Abstract

The present invention relates to method for communicating between a plurality of mobile nodes (1, 2, 3) in an ad-hoc network comprising the steps of: starting an application by a node (S31), wherein each application has assigned a preset priority value (Pappl); determining the amount of different applications running in the predetermined area (S34); calculating a sum of preset priority values in accordance to the amount of applications already running (S36); calculating a share of bandwidth depending on the preset priority value (Pappl) of the application and the sum of preset priority values; adjusting a network parameter according to the share of bandwidth of the application (S37); communicating by use of adjusted network parameter (S38). It further relates to an inter-vehicle network and a node within such inter-vehicle network. The invention provides a solution in which a total bandwidth is shared between all applications running at the same time in a predetermined area. Thus it is possible to implement a distributed management system providing quality of service to the nodes depending on the priority of the applications. By using an environment factor used for weighting and the recovery strategy, successful operation of applications having a low priority is achieved.

Description

Method for communicating between a plurality of nodes using a wireless communication protocol, Inter- vehicle network

The present invention relates to wireless ad hoc network and in particular to car-to-car communication, also called inter- vehicle network, in which sensor- equipped cars communicate cooperatively to e.g. avoid collisions or exchange information. Likewise car-to-car communication is considered crucial to avoid collisions during lane change/merge maneuvers and for reporting of invisible obstacles, possibly talking to obscured or shadowed objects.

An ad hoc network used as communication system provides a communication platform on top of which many different applications may run at the same time, each of them providing support to the user in different situations. In the particular application of car- to-car communication it is used to assist the driver for e.g. intersection assistance, rear-end collision avoidance, lane merging assistance.

Every type of communication equipment used for communicating is however limited in the amount of bandwidth it can supply in general and in particular to the applications running on the communication equipment. Every application requires a certain amount of bandwidth in order to work correctly, wherein the amount of bandwidth depends on the application and may be variable in time for each application. Moreover it is natural that among all the applications running, there may be some that are more important, because more directly related to the safe-driving context, and therefore need all the messages to be delivered with lowest possible delay. Other non-safety related applications e.g. peer-to-peer messaging or email exchange may have almost no delay constraints and their messages may be delivered using a best- effort mode. The WO 2005/074199 Al describes a method for improving communication between mobile nodes in an ad hoc wireless network to reduce the amount of data to be transmitted. It is described to form an application cluster and to select a cluster head managing the application cluster. Every application is able to organize its own application cluster, depending on the requirements of the application itself. For example a platoon of trucks following one another for a long time will constitute a platoon cluster to coordinate the speeds and directions of individual platoon members. The applications are free to access the wireless medium independently of each other. Thus it is not considered that in reality all the applications are part of one common communication system, and therefore they also have to share the system resources.

What is needed is a coordinated organization of all applications, where priorities of the applications play a key role.

Therefore it is an object of the present invention to provide a method for communicating between a plurality of nodes providing a quality of service for high priority applications.

The object is solved by the features of the independent claims. The invention bases on the thought that the bandwidth available for all applications within a predetermined area needs to be shared. In solutions know from the prior art the wireless medium is access independently from the total bandwidth. Thus critical delays may appear. According to the present invention it is possible to recognize how much bandwidth may be used by each application by determining the amount of applications running in the predetermined area and their demand of network resources. To provide a required quality of service for important applications like collision avoidance applications each application has assigned a priority value, which is considered by assigning a bandwidth share to be used by the respective application. Also in case of only running high priority applications the total bandwidth needs to be shared resulting in an assigned bandwidth share depending on the total available bandwidth and the priority value of the application. Thus the method according to the present invention provides a bandwidth assignment to running applications assuring the quality of service or throughput time for messages of the applications accessing the medium. Further the inventive method and the network system base on a completely distributed management, so no central controller is required that needs to decide the number and types of applications allowed to access the system in a given moment. The invention provides further an inter-vehicles communication system that can support different applications with different quality of service requirements. Service differentiation is achieved by defining different priority levels for the applications, where preset priority values are based on the type of application. The preset priority level basically defines the amount of bandwidth that can be reserved by an application. Based on the preset priority value a share of the total available bandwidth is derived and used. The assigned bandwidth share is used by adjusting network parameters used for communicating.

In order to implement such method concept, different applications are assigned different priorities. The assignment of priorities is application dependent. The priorities are preferably determined at the application level, where networking parameters are defined such as back off time, retry window size, data rate etc. The networking parameters may influence the communication behavior implemented in the lower protocol layer of the communication platform.

The system is completely distributed, so that there is no central controller that can decide the number and types of applications allowed to access the system in a given moment.

Further advantages will by described in respect to the dependent claims. According to a preferred embodiment a node is starting an application, wherein each application has assigned a predetermined or preset priority value. There are two situations, which may appear.

In a first situation the node that is starting the application is starting that application as first. No further node is running such application. In this case the node has to determine by itself how many other applications are running. This is achieved by polling the neighboring cluster heads or neighboring nodes. Each node knows its neighboring nodes by sending beacon signals at predetermined periods. After having polled the neighboring nodes the total amount of running applications is known to the requesting node. Now the share of bandwidth needs to be calculated by the node. The priority values of the other running applications may be requested by a further request, or they may be sent together with the answer received on the first request for determining the amount of different applications. Having calculated the bandwidth share depending on its own priority value and the sum of preset priority values of the other applications, predetermined network parameter like transmitting power or back off time may be adjusted to use the calculated bandwidth share for communicating.

In a second situation the application that is starting an application polls the neighboring nodes to determine whether such application is already running or whether such application cluster already exists. In case of already running such application the polling node is joining the application cluster. The selection of the cluster head is not the subject of the present invention. It may be selected arbitrarily among the involved nodes or according to predetermined characteristics.

Further the communication protocol used for communicating between the nodes may be chosen arbitrarily. Each application cluster is controlled by a cluster head, wherein the cluster heads are temporarily assigned and may be changed during the running of the application e.g. since the cluster head is leaving the transmission range or is closing the application.

In the second situation the cluster head is determining the amount of different applications within the predetermined area by polling the neighboring cluster heads. The cluster head is calculating the bandwidth share for its application cluster depending on the priority value of the application cluster and depending on the sum of preset priority values of the other applications. Depending on the calculated bandwidth share the cluster head is informing each node participating the application cluster to adjust its network parameter according to the bandwidth share for communicating. The preset priority value determines which percentage of the overall bandwidth the application layer is allowed to use: from the application point of view higher priority means higher share of total bandwidth available. This will determine the amount of information each application is allowed to generate for transmission. Internally, each node will divide its available bandwidth between the running applications depending on their respective priority number In a further preferred embodiment the preset priority value may be weighted by an environment characteristic factor resulting in a current priority. Applications may have different relevance. Their importance depends both on the type of application and on their current environment e.g. location of the vehicles, road and traffic conditions.

The environment characteristic factor represents an increase or decrease of the priority of the application. It can be either positive or negative of the initial preset priority value, depending on local information, like position information or other relevant data received from sensors or from other messages, as for example information about road traffic conditions. In this way the current priority is made dynamic, reflecting the fact that certain applications may become more important in specific zones or situations e.g. an "intersection assistance" application will have higher priority in the proximity of a big intersection, or a "Warning Dissemination Application" will have higher priority if ice on the road is detected and the road traffic is quite dense at that moment. In a preferred embodiment the current priority may be a function of expiration time and range because of the node mobility. When a predetermined time is expired or the vehicle is outside the range predefined by the application, the priority of the application is reset to the standard preset priority value.

Since the cluster head should react on changed environmental conditions it can at any time re-evaluate its priority level by changing the environment characteristic increase/decrease factor.

In a further preferred embodiment it is checked whether the assigned bandwidth share fulfils minimum requirements for maintaining the application. Every application has a minimum set of requirements that it has to satisfy: if medium or communication channel conditions do not permit to achieve this minimum requirements a recovery strategy depending on available bandwidth and priority is activated.

For determining the minimum requirements each node includes means for metering in order to monitor the bandwidth share per application. Corrective actions are taken in case the minimum requirements to sustain the applications cannot be met. In case of not sustaining the application, the cluster head is notified to increase the priority level. On the other hand it may preferable, if the cluster head does not receive such indications during a specified time interval, that the cluster head will automatically decrease the priority, so bandwidth can be reused by other more demanding applications. In both cases, all the cluster heads regularly exchange information about the updated priority list and recalculate their respective bandwidth share.

If the minimum requirements are not fulfilled the bandwidth derived from the distributed priority calculation is not enough to support the application. So the cluster head increases its priority number so that all the bandwidth share of the running applications is recalculated: in this way the application will get a higher part of bandwidth. The same thing can be done also by other applications, so in principle the procedure according to the present invention would start to oscillate until the steady state is reached where all applications are at the Priority level originally assigned by global standardization.

Preferably three stabilizing factors may be introduced:

There may be a range of priority values for an application, which can be requested (i.e. a maximum priority value per application type). Further by defining the rate at which an application can propose or request a higher priority value the time for finding a steady state may be controlled. The Rate may be defined also depending on the application type. This means that more important applications can more frequently request higher priority than lower priority applications.

A priority increase factor may be used which is dependent on the application types, meaning that higher priority applications can more quickly ramp up the bandwidth. This mechanism helps to satisfy the short term bandwidth demand of high priority applications, sacrificing quality of service for lower priority applications

If after a certain number of priority increase requests the bandwidth is not enough, the Cluster Head can send a message to request the other Cluster Heads to free some bandwidth. When this procedure is activated, all the running applications are evaluated and the one with lower priority is put in a 'sleep mode' for a certain amount of time. In this way more bandwidth is made available to the other applications. The applications in 'sleeping mode' are not allowed to send any message for a certain period of time. The time may also depend on the type of application, after which they start accessing the medium following again the whole procedure described above.

The adjustment of network parameters depending on the bandwidth share may include an increase or decrease of packet length in the upper layer. This may not influence the MAC or the physical layer. An increase of packet length means a higher bandwidth share is assigned. A decreased packet length means a lower bandwidth share. Further it is possible to adjust parameters like back off time for accessing the medium, wherein a shorter back off period means a higher probability to access the medium, thereby increasing the bandwidth share. Further parameter may be the repetition rate of resending a packet or the transmission power. Nodes having a high transmission power can more likely successfully transmit their data and therefore on average use a higher bandwidth share than nodes having a lower transmission power. The object of the invention is also solved by an inter- vehicle network comprising a plurality of vehicles and/or objects using the method for assigning bandwidth share as described above.

Further the object of the invention is also solved by a mobile node communicating in a wireless ad hoc network comprising a plurality of nodes. The node comprises processing means comprising means for starting an application, wherein each application has assigned a preset priority value (P_appi); further means for determining the amount different applications running in a predetermined area; the node is provided for calculating a sum of preset priority values (P_appi) in accordance of the amount of applications already running and for calculating a share of bandwidth depending on the preset priority value (P_appi) of the started application and the sum of preset priority values, wherein network parameter used for communicating are adjusted by the node according to the calculated share of bandwidth of the application.

The invention is described in detail below with reference to the accompanying schematic drawings, wherein

Fig. 1 shows a traffic situation having a plurality of application clusters;

Fig. 2a, b show pie charts illustrating calculated bandwidth shares;

Fig. 3 shows a flow chart illustrating the inventive method; Fig. 4 shows a further flow chart illustrating an alternative embodiment according to the inventive method. The drawings are provided for illustrative purpose only and do not necessarily represent practical examples of the present invention to scale.

In the following the various exemplary embodiments of the invention are described.

Although the present invention is applicable in a broad variety of applications it will be described with the focus put on car-to-car communication in an inter- vehicle network. A further field for applying the invention might be the assignment of bandwidth share to different applications in a general ad-hoc network having a plurality of notebooks or wireless sensors communicating via any type of wireless communication protocol.

Figure 1 illustrates a typical traffic situation on a road having multiple lanes. A first cluster includes the platoon of trucks 1. The common application controls the distance and the speed between the trucks over a long time. The cars 2 on the interweaving lane and the trucks 1 on the high way form a second application cluster. The common application realizes the distance and speed control for the interweaving cars 2. The cars 3 surrounding the accident form a third application cluster. The common application is warning the succeeding cars. A further application cluster may be used for exchanging weather information or traffic jam information. As will be easily noted the different applications have different importance. The accident on the middle lane has the highest importance. Thus the warning application for warning succeeding cars has a high preset priority. By approaching the location of the accident all cars will be warned by the warning application and will start to exchange sensor information with each other and/or the cluster head. Further applications having a high priority also are the interweaving application and the platoon application. Both are important, however the warning due to the accident is more important resulting in a slightly lower preset priority value for the application clusters "platoon" and "interweaving". The exchange of weather and traffic jam information is not that important as the application mentioned above. However the traffic information exchange is more important than the weather information. Further applications may be email exchange or information of local tourist attractions or hotels. Since the members of an application cluster changes dynamically it is important to support the message exchanges for applications having the highest priority.

Further not illustrated situation for different applications may be work zone warning, lane merging or intersection assist. Before explaining the invention by use of further embodiments the calculation of the current priority value and the calculation of the bandwidth share will be explained in more detail.

The current priority value P_tot depends on the preset priority value P_appi and the environmental factor I_env Each application as mentioned above has assigned a preset priority value. Due to the simplicity it will be explained using three different preset priority values. The warning applications, e.g. accident warning, work zone warning, are used to warn the approaching driver. They have the highest priority of 3.

The assisting applications, e.g. interweaving, platoon, intersection, lane merge, are used to support the driver only to increase the safety. The have the preset priority value of 2. The information exchange applications, local attraction, hotel, weather, email etc, will have the lowest priority of 1.

However all different applications are processed within the same predetermined area. Further a car may be executing different applications in parallel.

For example the warning application to indicate the accident may be executed at the same time as a information application for tourist attractions. An application may be started by a sensor of the car or by the driver. Further it is possible to start an application after having received a broadcast message containing a warning due to an accident in front of the car.

Since the conditions during driving may change it is important to have the possibility to weight the preset priorities P_appi of the applications. To weight the preset priority values P_appi an environment factor I_env is used. The environment factor

Ienv may be increased if the traffic density increases. If the traffic density decreases it may be decreased also. However in such cases the environment factor I_envwill be used to increase the preset priority value for highest (3) and high (2) priority applications only. Low priority (1) applications may be excluded from the weighting.

Considering the example above an increased traffic density will increase the environment factor I_env and thereby the current priority P_tot of the warning and assisting applications having a preset priority of 3 or 2. Thus they will be assigned more bandwidth than in the case without weighting by use of the environment factor I_env.

The following equation is used for calculating the current priority P_tot of a certain application. PtOt= Pappl Ienv (1)

Each application has assigned its preset priority P_appl. This may be a feature of the application, which is defined during installing the application within the vehicle. The environment factor I_env may be sensed by an application monitoring several sensors attached on the cars for checking the traffic density or the temperature etc. A central processing unit of the car performs the processing.

Since each application requires its bandwidth it is necessary to consider the total bandwidth available for communicating. Therefore the vehicle starting the application or the cluster head of an application cluster is asking neighboring cluster heads or vehicles which or how many applications they are performing at the moment. The requesting vehicle or cluster head will receive the current priorities P_tot from the neighboring vehicles or cluster heads. Having the own current priority value P_tot calculated by use of the equation (1) as explained above the bandwidth share B_part may be calculated. This is achieved by use of the following equation.

BandwidthShare = TotalBandwidth ■ ^{r ιonty}"" (2)

2,Pr iority_t'_ot i The total bandwidth available is a given feature, which is known to the inter- vehicle network.

The use of the equation (2) is explained on the following example. A vehicle is assigned to start an accident warning application. The predefined priority level for this highly safety relevant application is 3. The vehicle informs about other available application clusters in the neighborhood, and fills in a table listing all applications and corresponding priorities. At the same time it tells these cluster heads that it is going to consume part of the shared bandwidth for an application of priority=3.

The bandwidth share for the accident warning application in our example is 3/12 = 25%. All cluster heads acknowledge the receipt of this message, update their application cluster table and calculate the Bandwidth shares for their respective applications, which will obviously be lower than before. In practice, the priority is reflected in a percentage of bandwidth shares, which is reflected in a percentage of the standardized communication network parameters that have impact on the throughput, most obviously examples: - Message length (in our example: 25% of the standard 1KB)

Message repetition rate (in our example 25% of the standard 10Hz)

Other network parameters influencing the throughput and bandwidth usage are the number of parallel channels or the contention window size, where 25% of max value here means a 4-fold increase of the number of time slots to wait for accessing the medium again.

The system relies on a set of standard values for the basic network parameters constituting the bandwidth or throughput demand of each application. This set of parameters can vary from country to country and depends on the physical layer used, which is known to every vehicle before participating in the communication (e.g. 2.4GHz WLAN, or 5.4GHz WLAN or code division multiplex system etc.)

In the following a table 1 is shown illustrating different applications having the ID 1-7.

Table 1 According to the priority value given in table 1 above the bandwidth shares are calculated. The resulting bandwidth distribution is illustrated in the pie chart in figure 2a. Thus the application cluster having ID 1 running the application accident avoidance is getting the highest bandwidth share of 25%. The applications having lower priority e.g. weather info, will get a bandwidth share of 8% only, since it may accept higher delays during accessing the medium than the accident avoidance application.

By changing the priority of an application due to the weighting factor I_env the total bandwidth stays the same. However an increase of a bandwidth share of a certain application will decrease the bandwidth share of the other applications running at the same time. This is illustrated in table 2 and the respective pie chart in figure 2b.

Table 2 In the illustrated example the preset priority P_appi of the accident avoidance application is weighted by the environment factor 1,3 resulting in a current priority value P_tot of 5 and in a bandwidth share of 36%. The bandwidth share of the other application is decreased to 14% or 7% respectively.

As mentioned above the functioning of an application depends on a bandwidth share. If the bandwidth share is to low the application does not get the required data in due time to assure error-free transmission. To avoid such situations the minimum requirements are controlled by a metering function in each vehicle. The metering function may measure how much data is transmitted or received within a certain period of time. Preset values for the amount of transmission and receiving are stored in the vehicles. Thus a comparison will be performed to indicate whether the minimum requirements are fulfilled. I case of a bandwidth share being to low a recovery strategy is started. The cluster head is then increasing its priority value, wherein this priority value is transmitted to the other cluster heads. They have to recalculate its bandwidth share due to the increase priority value of the requesting cluster head. As mentioned above several steps are possible to handle such recovery strategy.

In the following the flow chart of figures 3 and 4 are described. In Step 31 a vehicle starts an application and is polling neighboring nodes whether such application is already running in step 32. In case that such application is not running the node itself will act as cluster head. In step 34 it will request from neighboring cluster heads how many different applications are running at the same time. The node will receive the priority values of the neighboring cluster heads. In step 35 the own current priority P_tot is calculated reflecting different environment situations. The bandwidth share which can be used by the application is calculated in step 36. The network parameters are adjusted according to the calculated bandwidth share. The node will communicate by using the bandwidth share. Thus a quality of service for this application having a guaranteed throughput is provided.

In case that such application is already running the requesting node is joining the application cluster in step 41. This may be the case in the traffic situation of figure 1. A car is approaching the accident. It will receive a warning message. This message will start the accident avoidance application having the highest priority. In step 42 a cluster head will be selected. This can be done by known procedures. For example the vehicle staying the longest time in the application cluster will be selected. This may be the car, which is broken down. The cluster head is requesting the other cluster heads how many different application are running. It will receive the priority values of the other cluster heads, e.g. from the weather info application cluster etc. Then in step 44 its own current priority P_tot is calculated considering the environment factor I_env and the preset priority value P_appi. In the following step 45 the bandwidth share for the application cluster "accident avoidance" is calculated. The bandwidth share may be divided equally among the participating vehicle in the application cluster. It is possible also to assign the bandwidth share of the application cluster to all vehicles of the application cluster. The contention between the vehicles may be handled by known contention processing method. The calculated bandwidth share is realized by an adjustment of the network parameter, e.g. the packet length. Thus an application using a higher packet length due to its higher bandwidth share will access the medium with higher throughput. The present invention provides a solution in which a total bandwidth is shared between all applications running at the same time in a predetermined area. Thus it is possible to implement a distributed management system providing a quality of service to the nodes depending on the priority of the applications. By using the environment factor and the recovery strategy, successful operation of applications having a low priority is possible.

Claims

CLAIMS:

1. Method for communicating between a plurality of mobile nodes in an ad-hoc network comprising the steps of: starting an application by a node, wherein each application has assigned a preset priority value (P_appi); - determining the amount of different applications running in the predetermined area; calculating a sum of preset priority values in accordance to the amount of applications already running; calculating a share of bandwidth depending on the preset priority value (Pappi) of the application and the sum of preset priority values; adjusting a network parameter according to the share of bandwidth of the application; communicating by use of adjusted network parameter.

2. Method according to claim 1, wherein the step of determining the amount of applications comprising further: polling other nodes to determine whether such application is already running within a predetermined area to join an existing application cluster; selecting a cluster head among the nodes running the application; - polling neighboring cluster heads for determining the amount of applications.

3. Method according to claim 1, comprising further weighting the preset priority value (P_appi) by an environment characteristic (I_env) resulting in a resulting current priority (P_tot).

4. Method according to claim 1 to 3, comprising further monitoring within each node whether the assigned share of bandwidth of the application is sufficient to maintain the quality of service required for the application.

5. Method according to claim 4, further comprising: in case of not maintaining the functionality of the application: starting a recovery strategy by transmitting an increase request to the assigned cluster head or in case being itself the cluster head to the neighboring cluster heads.

6. Method according to one of the preceding claims, wherein the network parameters include at least one of channel packet length, repetition rate, transmit power, back off time, retry window size, data rate.

7. Inter- vehicle Network including a plurality of vehicle and/or objects communicating via a wireless communication protocol using the method as claimed in one of the claims 1-6.

8. Mobile node communicating in a wireless ad hoc network comprising a plurality of nodes having means for starting a application, wherein each application has assigned a preset priority value (P_appi); further means for determining the amount of different applications running in a predetermined area; the node is provided for calculating a sum of preset priority values (P_appi) in accordance of the amount of applications already running and for calculating a share of bandwidth depending on the preset priority value (P_appi) of the started application and the sum of preset priority values, wherein network parameter used for communicating are adjusted by the node according to the calculated share of bandwidth of the application.