外文翻譯.pdf

SICE Annual Conference in Fukui. August 4-6,2003 Fukui University, Japan Implementation of Inter-Vehicle Communication System for Vehicle Platoon Experiments via Testbed Tae Min Kim, Jae Weon Choi School of Mechanical Engineering and Research Institute of Mechanical Technology Pusan National University, Pusan 609-735,Korea tmkim,choijuQpusan.ac.!u Abstract: This study considers the implementation issues of the inter-vehicle communication system for the vehicle platoon experiments via a testbed. The testbed, which consists of three scale vehicles and one RCS(remote control sbation), is developed as a tool for functions evaluation between simulation studies and full-sized vehicle researches in the previous study. The cooperative communication of the vehicle-to-vehicle or the vehicle-to-roadside plays a key role for keeping the relative spacing of vehicles small in a vehicle platoon. Then: the traffic capacity could be increased greatly. The static platoon control, where the number of vehicles remains constant, is sufficient for the information to he transmitted in the suitably fixed interval, while the dynamic platoon control such as merge or split requires more flexible network architecture for the dynamical coordination of the communication sequence. In this study, the wireless communication device and the reliable protocol are implemented for the flexible network architecture, using the lowcost, short-range, IShI band transceiver and the 6-bit microcontroller. Keywords: vehicle platoon, wireless communication system, testbed 1. Introduction Urban road in most major cities becomes congested more and more because the demand of travel exceeds highway capacity. The congestion problem causes many other problems: the waste of time and energy, the traf- fic accident, the pollution, and so on. ITS (Intelligent Tkansportation System) is developed actively as a SD lution of these problems. Especially, IVHS (Intelligent Vehicle and Highway System) is the major subject in ITS. The purpose of IVHS is to improve safety as well as l,o increase a highvmy capacity through automated vehicles and automated highways.) In IVHS, a exem- plary method of efficient vehicle control by grouping in plat.oons has been proposed in PATH program.) The vehicle platoon is a group of vehicles traveling together at a high speed with relatively small spacing. Vehicles in close-formation platoons are dynamically coupled hy feedback control laws. Depending on the information available for feedback and depending on how such a11 in- formation is processed in the synthesis of an automatic vehicle following control law, dynamic interactions be- tween vehicles can cause instability in a vehicle string. The control with the information of the lead vehicle which is the first vehicle in a platoon and the preced- cle. For the static platoon control, where the number of vehicles remains constant, it is sufficient, for the iofor- mation to he transmitted in the suitably fixed interval since each vehicle dose not require the frequent update of the control input. This scheme guarantees that each vehicle in the platoon has an opport,unity to transmit its information every one cycle. For the dynamic pla- toon control, such as merge or split, the more flexible network architecture is required. In this case, since the maneuvering vehicle for merge or split requires the fre- quent update of the control input, the more inforniatioo should be transmitted to the maneuvering vehicle than others. Therefore, the more opportunity to transmit the information is given to the lead vehicle and the ma- neuvering vehicle. In this scheme, the cominunicat.ion sequence should he coordinated effectively. The coordi- nation of the communication sequence may be achieved easily by the RCS(remote control station). In this study, the wireless communication system, which can coordinate the communication sequence by the RCS- is implemented for the vehicle longitudinal platoon experiments. 2. System Requirements ing vehicle can only guarantee the stability in a vehicle striig.) The information of the preceding vehicle can be ob- tained relatively by the range radar. But the informa- tion of the lead vehicle is not available to all vehicles in a platoon. The wireless communication only enables all vehicles to obtain the information of the lead whi- n&jC flow capacities are affected directly vehicle pla- toon control strategies. The effectiveness of a platoon control strategy can be gauged by the maximum traffic flow capacity, the attenuation of spacing errors that it can guarantee, and the amount of information that is needed to implement the strategy in real-time. There are twn main methods of control that have been studied 3414 PR0001/03/0000-1823 $400 0 2003 SICE Remte Canlrol Scation ? . El Fig. 1: The configuration of the testhed Fig. 2: The configuration of a scale vehicle of the testbed in IVHS: comtant spacing and constant headay.) In constant spacing control, the desired spacing between vehicles is tracked whereas in constant headway con- trol, a desired headway which is the time it takes a vehicle to cover the distance between itself and the pre- ceding vehicle is maintained. The advantages of using constant spacing over constant headway control is to increase the throughput of vehicles on the highvay, al- though constant headway control is more favorable since no external information is required. In constant spacing control, external information is required for string sta, bility. A wireless communication system can be utilized to transfer this external information. In vehicle platoon system with constant spacing strat- egy, the following vehicle needs the information of the preceding vehicle and the information of the lead vehi- cle. The information includes the position, the velocity, the acceleration, and the particular command of the preceding or the lead vehicle. At the same time, each vehicle may need the command of a RCS due to a spec- ified event such as emergency. In the previous study4), the testhed for vehicle longitudinal platoon experiments is developed. The testbed shown in Fig. 1 consists of three scale vehicle and a RCS. Each vehicle includes, as shown in Fig. 2, the sensors for the data acquisition of available information, the operating system for the com- putation of the control command, the actuators for the driving and steering for the most effective maneuvering, the wireless communication system for the exchange of external information, and the interface for the synthesis of basic functions. In the previous stndy4), the 433hIHz RF-module, BIAJ-433, is used for the implement,ation of the wire- less communication system of which the architecture is the TDhIA (Time Division lZIultiple Access) with to ken passing. The data transfer rate of this Id-module is 38Kbps and the carrier sense algorithm is not sup- ported. Therefore, the performance of the communica- tion system is not sufficient for the sequence scheduling algorithm5. G, and the control synchronization with ve- hicles. For stable movement of each vehicle, the sam- pling period of a vehicle should be less than 401s. Sen- sors used in testbed can be satisfied with the sampling period of 30ms. It takes above 5ms to transmit the data (12bytes) and the preamble (above 3ms) at 38Kbps. To improve the performance and rohnstness of the com- munication system, more time is needed because the overhead of packet is increased. In addition, the verifi- cation with a variety of network sequence scheduling al- gorithm wants more flexible network architecture. The wireless communication system is composed of both the hardware and the software. The hardware provides the means to connect various stations (vehicles or RCS) in the network while software provides smart control over the hardware components. The software will also pro vide a flexible and reliable protocol for the exchange of communication data. 3. Hardware Implementation The hardware of the wireless communication system is composed of the following components: t,he RF front- end module, the int.erface chipset reserved for RF front- end module, and t,he hICU(microcontrol1er unit). The Fig. 3 illustrates the configuration of the wireless com- munication system. The architecture of the hardware is classified into four layers as follorvs: PHY layer (Physical layer) PHY-RIAC layer (Physical-to-MAC layer) MAC layer (Medium Access Control layer) MAC-APP layer (AIAC-twApplication layer) The PHY layer is achieved by the RF front-end module, RFWI02 transceiver (transmitter/receiver) chipset which is developed by RFUaves Ltd.) The PHY-hIAC layer is constructed by the interface chipset, the IJO (input/outpnt) ports of AICU, and the IJO driver in AICU. The exclusive interface chipset, the RFWD100 developed by RFWaves Ltd., is used for the interface of the RFW-102.8) The hIAC layer, which is a set of protocols for maintaining order in the use of a shared medium, is the software in AICU. The AIAC layer is discussed in later chapter because it is the soft- ware component. The MAC-APP layer is the interface part between the wireless communication system and the vehicle or RCS. 3415 PHI-IMC Fig. 3: The configuration of the wireless communication system Operationg frequency Serial UART Memory Table 1: The specification of the RFW-102 Transfer rte I 11 tn IMhnn Physical Media I DSSS, ISM Band (2.4GHz) 3.6864hIHz 2EA(up to liblbps) 16Kbytes(flash), -. -. . _. Ir ._ _._. Bandwidth I 30hlhz at -20dB Timer/Connter 8-bit(ZEA), lB-hit(lEA) 3.1 PHY layer The PHY layer actually handles the transmission of dat;a between the wireless communication system. In this layer, RFW-102 transceiver chipset is used BS the RF front-end module. The motivation for the RFW-101 is its high data transfer rate, the ease in interfacing to an external de- vice, the availability for carrier sense algorithm, and the inexpensiveness. Table. 1 shows the specification of the RFTV-102. Since the transceiver chipset provides a peak output power of 2dBm and the sensitivity is -80dBm when BER (Bit Error Rate) is the transmission is available up t,o 30m in open. This range is suitable for the testbed using scale vehicles of which the size is about 0.3m. 3.2 PHY-MAC layer PHY-hJAC layer is the interface between the RF front- end module and AIAC protocol. The layer is con- structed with the following components: the interface chipset reserved for RF front-end module, the 110 ports of the microcontroller unit, the driver for 110 ports. The RFW-D100, developed by the RFIVaves Ltd., is used as the interface chip. The RFW-D100 is a com- plimentary chip to the RFW-102 chipset. It proxrides a parallel interface to the RFW-102, and other features which make it easy to implenient. a protocol suitable for wireless communication. For this study, the hICU is in charge of the MAC layer protocol and the driver for 110 control. The interface chip reduces the real-time demands of the AlCU handling the MAC protocol. The interface chip gives the hICU an easy parallel interface with the RF front-end module, similar to memory ac- cess. The interface chip converts the fast serial input from the RF front-end chipset to 8-hit words, which are suitable for an 8-bit hICU to work with. In addition, the interface chip requires a lower rate oscillator for idle mode. In the idle mode, the power consumption of the BER 3416 at -80dBm I lKbytes(SRAh1) External Interrupts I 3EA RFW-102 and the FWW-0100 is greatly decreased. The interface chip buffers the data through a l6byte FIFO (First In First Out buffer), which is giving the hlCU access to the RFW- DlOO more efficiently. Instead of reading 1 byte per int,errupt, the hICU can read up to 16bytes in each interrupt. In cases where each incoming byte causes an interrupt, this reduces the overhead of the MCU in reading incoming words, insofar as stack stuffing and pipeline emptying are concerned. When using the FIFO, the AICU pays the same overhead for all the FIFO bytes, as it paid for only one byte without a FIFO. The hICU handles actually the RFIV-D100 by AIAC protocol and application. The ATmegalGlL, developed by ATAIEL Corp., is used as the AICU.g) Table 2 shows the specification of the ATmegalGIL. In this study, the ATmegalGlL allows two external interrupts to be invoked by the RFWD100. The state of MAC layer is changed by the interrupts invoked by the event of the RFW-D100. According to the change of the state, MAC layer executes particular functions, such as receiving, transmitting, error check, acknowl- edgement, and other handling of data. 3.3 MAC-APP layer MAC-APP kyer is the interface between MAC layer and APP(application) layer. In this study, the applica- tion layer is the control loop of each vehicle. The APP layer is connected to the programmable serial UART, which has one interrupt vector. Therefofe, the inter- rupt invoked by the application layer also coordinates the state of MAC layer and the particular functions are achieved. In addition, the redundancy of internal SRAhI is allocated for the receiving and transmitting buffer, which extends the FIFO in WWD100. Then, it becomes possible to transmit and receive longer size of data than FIFO in RFW- DZOO continually. 4. Software Implementation Fig. 4 illustrates the block diagram of software configu- ration in this study. Software configuration is classified into three layers as follows: PHY-MAC layei MAC-APP layer MK-APP Fig. 4 The block diagram of software configuration MAC layer The hIAC layer is also divided into hIAC state and MAC data. The general procedure of the XIAC protocol is as follow!;: 1. The external interrupt invoked by the PHY layer or the ,UART interrupt invoked by the APP layer activates the MAC state management. 2. The hlAC state management checks the stat,us of the PHY layer or the APP and changes the hIAC state in accordance with the result of that. 3. According to the newly changed MAC state, the MAC ,data management controls the data stream and the RX/TX buffer. 4. And then, in accordance with the result of the RIAC data management, the MAC state manage- ment sets the new MAC state. Two external interrupts of the hICU is allocated for the PHY layer and UART interrupt is allocat.ed for the APP layer. The transition of the MAC state is driven by these interiupts and the bbit timer controls the execu- tion time. Therefore, the configuration of the protocol logic is improved effectively and the real-time execution is guaranteed. In addition, RX buffer and TX buffer in the MAC 1:Lyer have 64hytes of SRAII, respectively, for the constraint of the FIFO size in the RFWD100. 4.1 Reliiable protocol In genera1,lthe communication is done by the exchange of the packet which is a set of data for efficiency and conveniencs:. In this study, a packet, as shown Fig. 5, consists of the following fields: Preamble: to synchronize the receiver side Fig. 5: The configuration of a data packet for vehicle to vehicle Network ID: to filter a packet from other network Destination: Destination ID Source: source ID Type/Seq: Type of packet/Numher of sequence Size: Size of the whole packet DATA: Actual data to transfer CRC: 16bit-CRC to check the validity of the packet The data field includes the position, the velocity, the acceleration, for the transmission of each vehicles data. For the command packet of the RCS; the data field in- cludes the comma.nd of t,he RCS. In this study, the IS11 band RF transceiver is used. Since IShI band is a shared resource between many wire- less applications such as IEEE 802.11 and Bluetooth, an overlap in time, frequency and space domain must oc- cur and the network may experience interference from other network. Each of the standards such as IEEE 802.11 and Bluetooth uses a packet-oriented protocol and utilizes the shared channel only for fragments of time. A protocol for an ISM hand application will have to use the time intervals for which the channel is free or relatively free (the interference is weaker) in order to transfer the required data. When a node wants to transmit, the node listens to the channel and checks if the channel is free. This is done hy CS (Carrier Sense) mechanism supported by RFW-DlOO.) In order to make sure that a data packet arrives suc- cessfully at t.he destination (receiver), the source (trans- mitter) needs to get some verification from the receiver side within a certain fixed time. The transmitter will get this verificat,ion, by getting an acknowledge packet from the receiver side. If the transmitter does not get an acknowledge packet, the transmitter tries to retransmit the data packet. 4.2 Protocol behavior For the static platoon control, since each vehicle dose not need the frequent update of the control input, the chance of the transmission of the data packet is allo- cated uniformly in each vehicle. For the dynamic pla- toon control, since the maneuvering vehicle for merge or split requires the frequent update of the control in- put, the more information should be transmitted to the maneuvering vehicle than others. This is the reason 3417 ,. I- I* I (201 m mr.o*.* m *a* I/om.n/,* Tam, CCm Fig. 6: The example of the protocol behavior wby the communication sequence is coordinated dy- namically. The algorithm of finding the communication sequence was worked in Choi and Fangs work.5: The coordination of the communication sequence is achieved by the RCS. Fig. 6 shows the example of the prot.oco1 behavior in one cycle between wireless communication systems of this study. The communication sequence is #I, #2, #R. The major feature is that the RCS broadcasts t,he communication sequence command packet and the data update command packet every one cycle. At first, the RCS broadcasts the communication sequence com- mand packet. Then, each vehicle tries to transmit the data packet according 1.0 the communication sequence command. For the synchronization of t