基于ARM9和嵌入式Linux的车载障碍物自动报警系统
|
摘要
车载障碍物报警系统是车辆智能交通系统的一个重要课题,为驾驶员提供了一个安全、有效的驾驶环境。本系统的主要功能为用车载摄像头采集道路视频信息,并利用图像处理算法,对本车所在车道的左右分道线和阻碍本车行驶的障碍物进行检测。当在规定距离范围内检测到障碍物时,通过声音报警器和光学报警器,从声、光两方面提醒驾驶员注意前方障碍物;当障碍物距离本车小于临界安全距离时,即有可能与本车发生碰撞时,通过电子制动装置,使汽车强制停车,防止撞车等意外的发生。
根据项目需求、结合车载报警系统的发展现状,同时考虑成本等因素进行系统架构。设计并实现了ARM主控节点与CAN网络通讯模块的硬件电路以及基于CAN网络的各执行机构的硬件电路;基于硬件电路开发了基于MCP2510的CAN设备驱动程序和基于CAN网络的声音报警模块、光学报警模块以及电子制动模块的软件驱动;对采集到的道路图像进行处理,识别车道以及车道内的障碍物。根据障碍物的等级,指导相应执行机构工作。
本系统将道路图像处理与CAN总线网络结合起来,系统可根据障碍物的等级指导相应执行机构工作,具有声、光报警以及电子制动的功能,在一定程度上实现了车辆辅助安全驾驶的功能。
With the development of the urbanization and prevalence of vehicle, transportation problems are getting worse and worse. In order to solve the problem of heavy traffic and increasing vehicle accidents, intelligent transportation system is introduced. As an important part of ITS, the barriers’automatic alarm system on vehicle provides drivers with a useful and practical assistant safe driving system. According to statistics, unsafe distance and fatigue driving are the immediate cause of our country’s highway accidents. As shown in an European research, 50% accidents can be avoided by a five minutes early alarm to divers before the collision, while 90% accidents avoided by one minute early alarm. Obviously, it is very necessary to introduce barriers’alarm system into driving process.
Barriers’alarm system is generally composed of vehicular sensors (microwave radar, laser radar, video camera and other kinds of sensor), vehicular computer and execution mechanism and so on. Firstly, the alarm system measures objects around by vehicular sensors and finds out barriers which block the car or endanger the driver. Then, vehicular computer processes the data according to the barrier situation. At last, the execution mechanism sends corresponding orders to realize the alarm function.
With a reduction in the cost of optical devices and embedded computers, intelligent vehicle will take the responsibility of monitoring drivers and road condition in the next couple of years. Meanwhile, with the economic development in China, more families can afford for vehicle, which results in a huge vehicle consumer market. Therefore, barriers’automatic alarm system will have a very broad market prospects. Currently, the United States, Japan, Germany, Italy and other developed countries’research institutes have some achievement in this field. Relatively, our country has a certain gap compared with those developed countries. Now some our country’s institutes are working on this field, such as Jilin University, the National Defense University and Tsinghua University.
This paper is generated on the project of State Development and Reform Commission, named as Research and Industrialization Project on Vehicular Information System Based on IPV6. The main functions of this system cover acquiring real-time road image by the vehicular video camera and detecting the
right, left lane divider and the barrier in the lane by image processing algorithm. Additionally, when some barriers are detected within the specified distance, the sound and light alarm signals are sent to drivers, so the driver can slow down or bypass the barrier to avoid accidents, such as the rear-end collision and so on. If the barrier is so near as to nearly have collision with vehicle, electronic braking device is used to force vehicle deceleration or stop-off so as to prevent accidents and realize assistant safe driving.
Based on function analysis above, the main work of this paper is described as follows:
At the aspect of overall system design, requirement analysis and feasibility analysis go first according to project requirement. With a consideration of cost and the current development of barriers’alarm system, the program is decided to build on the basis of ARM9 and embedded Linux.
At the aspect of hardware, each hardware electronic circuit of modules is designed and implemented, including ARM and CAN network communication module and CAN Network module. The main steps are as follows: Firstly, according to selected chips’function, price and availability, each chip’s manual is comprehended and the parameters and the interfaces are obtained to confirm whether they are suitable for the system. Secondly, the cooperative capability of each chip is analyzed, in accordance with CAN Bus norms, ARM and CAN network communication module’s features and selected chips’features, then design all network nodes and develop chip communication parameters. Thirdly, the minimum circuit for each chip and the circuit for each node are designed. At last, the testing codes are programmed according to its basic function, and the basic hardware adjustment is completed.
At the aspect of software, each software function modules are designed and implemented as follows.
(1)Video input module: the main function is acquiring video information from road.
(2)ARM master control module: the core module of the system. Including the following sub modules:
Video image acquisition module, acquiring image from the video input module and saving image information in the designated buffer; video image Abstract
processing module, processing the video images acquired, identifying the right and left lane divider and barriers in the lane, and classifying barrier; execution module, using ARM and CAN network communication module to guide the corresponding part to take action according to barrier’s rank; ARM and CAN network communication module, carrying out the communication work with the CAN network through CAN controllers connected by SPI interface.
(3)CAN network execution module include three sub modules, including sound alarm module and optical alarm module, which produce sound alarm and light flickering to let the driver notice the barrier in front and take the action of a corresponding slowdown or bypassing the barrier to keep traffic safety, and electronic brake module, which is used to brake and decelerate vehicle to prevent accidents happened.
The innovations of the barriers’automatic alarm system are as follows:
1. The system uses the electronic network to integrate the inner structure of the vehicle and has a good prospect of application.
2. As the image acquisition, processing and vehicular CAN network are combined together, system can guide the execution module by different barrier situation, and realize safe assistant driving.
3. A new lane divider search algorithm is proposed. The new algorithm has characteristics of simple and faster processing speed, which are validated by final system experiment.
Due to time limitation, the system still needs some improvements. For example, this system is not suitable for barrier detection in a night driving condition. The following work is to realize the system’s real-time function in a real vehicle environment and connect system modules with corresponding nodes on vehicle to implement the wireless function.
根据项目需求、结合车载报警系统的发展现状,同时考虑成本等因素进行系统架构。设计并实现了ARM主控节点与CAN网络通讯模块的硬件电路以及基于CAN网络的各执行机构的硬件电路;基于硬件电路开发了基于MCP2510的CAN设备驱动程序和基于CAN网络的声音报警模块、光学报警模块以及电子制动模块的软件驱动;对采集到的道路图像进行处理,识别车道以及车道内的障碍物。根据障碍物的等级,指导相应执行机构工作。
本系统将道路图像处理与CAN总线网络结合起来,系统可根据障碍物的等级指导相应执行机构工作,具有声、光报警以及电子制动的功能,在一定程度上实现了车辆辅助安全驾驶的功能。
With the development of the urbanization and prevalence of vehicle, transportation problems are getting worse and worse. In order to solve the problem of heavy traffic and increasing vehicle accidents, intelligent transportation system is introduced. As an important part of ITS, the barriers’automatic alarm system on vehicle provides drivers with a useful and practical assistant safe driving system. According to statistics, unsafe distance and fatigue driving are the immediate cause of our country’s highway accidents. As shown in an European research, 50% accidents can be avoided by a five minutes early alarm to divers before the collision, while 90% accidents avoided by one minute early alarm. Obviously, it is very necessary to introduce barriers’alarm system into driving process.
Barriers’alarm system is generally composed of vehicular sensors (microwave radar, laser radar, video camera and other kinds of sensor), vehicular computer and execution mechanism and so on. Firstly, the alarm system measures objects around by vehicular sensors and finds out barriers which block the car or endanger the driver. Then, vehicular computer processes the data according to the barrier situation. At last, the execution mechanism sends corresponding orders to realize the alarm function.
With a reduction in the cost of optical devices and embedded computers, intelligent vehicle will take the responsibility of monitoring drivers and road condition in the next couple of years. Meanwhile, with the economic development in China, more families can afford for vehicle, which results in a huge vehicle consumer market. Therefore, barriers’automatic alarm system will have a very broad market prospects. Currently, the United States, Japan, Germany, Italy and other developed countries’research institutes have some achievement in this field. Relatively, our country has a certain gap compared with those developed countries. Now some our country’s institutes are working on this field, such as Jilin University, the National Defense University and Tsinghua University.
This paper is generated on the project of State Development and Reform Commission, named as Research and Industrialization Project on Vehicular Information System Based on IPV6. The main functions of this system cover acquiring real-time road image by the vehicular video camera and detecting the
right, left lane divider and the barrier in the lane by image processing algorithm. Additionally, when some barriers are detected within the specified distance, the sound and light alarm signals are sent to drivers, so the driver can slow down or bypass the barrier to avoid accidents, such as the rear-end collision and so on. If the barrier is so near as to nearly have collision with vehicle, electronic braking device is used to force vehicle deceleration or stop-off so as to prevent accidents and realize assistant safe driving.
Based on function analysis above, the main work of this paper is described as follows:
At the aspect of overall system design, requirement analysis and feasibility analysis go first according to project requirement. With a consideration of cost and the current development of barriers’alarm system, the program is decided to build on the basis of ARM9 and embedded Linux.
At the aspect of hardware, each hardware electronic circuit of modules is designed and implemented, including ARM and CAN network communication module and CAN Network module. The main steps are as follows: Firstly, according to selected chips’function, price and availability, each chip’s manual is comprehended and the parameters and the interfaces are obtained to confirm whether they are suitable for the system. Secondly, the cooperative capability of each chip is analyzed, in accordance with CAN Bus norms, ARM and CAN network communication module’s features and selected chips’features, then design all network nodes and develop chip communication parameters. Thirdly, the minimum circuit for each chip and the circuit for each node are designed. At last, the testing codes are programmed according to its basic function, and the basic hardware adjustment is completed.
At the aspect of software, each software function modules are designed and implemented as follows.
(1)Video input module: the main function is acquiring video information from road.
(2)ARM master control module: the core module of the system. Including the following sub modules:
Video image acquisition module, acquiring image from the video input module and saving image information in the designated buffer; video image Abstract
processing module, processing the video images acquired, identifying the right and left lane divider and barriers in the lane, and classifying barrier; execution module, using ARM and CAN network communication module to guide the corresponding part to take action according to barrier’s rank; ARM and CAN network communication module, carrying out the communication work with the CAN network through CAN controllers connected by SPI interface.
(3)CAN network execution module include three sub modules, including sound alarm module and optical alarm module, which produce sound alarm and light flickering to let the driver notice the barrier in front and take the action of a corresponding slowdown or bypassing the barrier to keep traffic safety, and electronic brake module, which is used to brake and decelerate vehicle to prevent accidents happened.
The innovations of the barriers’automatic alarm system are as follows:
1. The system uses the electronic network to integrate the inner structure of the vehicle and has a good prospect of application.
2. As the image acquisition, processing and vehicular CAN network are combined together, system can guide the execution module by different barrier situation, and realize safe assistant driving.
3. A new lane divider search algorithm is proposed. The new algorithm has characteristics of simple and faster processing speed, which are validated by final system experiment.
Due to time limitation, the system still needs some improvements. For example, this system is not suitable for barrier detection in a night driving condition. The following work is to realize the system’s real-time function in a real vehicle environment and connect system modules with corresponding nodes on vehicle to implement the wireless function.
引文
[1] 刘铁.智能驾驶中视觉导航算法及其 DSP 系统实现.硕士学位论文.2004:1-5.
[2] Broggi A, Bertozzi M, Fascioli A, ConteG Automatic Vehicle Guidance. The Experience of the ARGO Automous Vehicle.World Scientific, Singapore, 1999.
[3] 胡永裕.怎样消除高速公路危险源.现代职业安全,2005,(7):56-57.
[4] 鲍吉龙,应延治等.基于 DSP 技术的汽车防撞雷达.微计算机信息, 2006,22(4-2):188-190.
[5] C.Thorpe.Vision and navigation: The Carnegie Mellon Navlab. Kluwer Acade- mic Publishers, 1990:83-93.
[6] L.Andreone, A.Broggi.Vehicle Detection and Localization in Infra-RedImages. The IEEE 5th International Conference on Intelligent Transportation System, Singapore, September 2002:3-6.
[7] 徐友春 . 智能车辆视觉与 GPS 综合导航方法的研究 . 博士毕业论文.2001:1-10.
[8] 杜春雷.ARM 体系结构与编程.清华大学出版社,2003.2:1-2.
[9] Peter Hank,Egon J?hnk. SJA1000 Stand-alone CAN controller. Application Note AN97076.Philips Semiconductors, December 1997.
[10] 朱华旭.CAN/LIN 总线技术在车内网络的应用研究.武汉理工大学学报, 2005,5:210-212.
[11] 饶运涛,邹继军,郑勇芸.现场总线 CAN 原理与应用技术.北京航空航天大学出版社, 2003.6:14-19.
[12] User’s Manual S3C2410X 32-bit RISC Microprocessor Revision 1.2,Samsung Electronic,2003.
[13] 求是科技.单片机典型模块设计实例导航.人民邮电出版社,2004.5:349-421.
[14] 王巍,秦贵和,范玮.基于 CAN 总线的汽车车身线束设计.制造业自动化,2005,27(7):71-74.
[15] Peter Hank, Egon J?hnk, SJA1000 Stand-alone CAN controller Application Note, Philips Semiconductors, 1997.
[16] Eisele, H. and J?hnk, E.PCA82C250/251 CAN Transceiver. Application Note AN96116. Philips Semiconductors, 1996.10.
[17] Data Sheet MCP2510, Microchip Technology, 1999.
[18] 潘巨龙,黄宁,姚伏天.ARM9 嵌入式 Linux 系统构建与应用.北京航空航天大学出版社,2006.8:12-13.
[19] AT89S52 Datasheets.ATMEL corporation.2001.
[20] 7805-LINEAR INTEGRATED CIRCUITS 3-TERMINAL VOLATGE REGUL–ATORS ETC.http://www.alldatasheet.com.
[21] VNH3SP30 - FULLY INTEGRATED H-BRIDGE MOTOR DRIVER. http:// www.alldatasheet.com.
[22] 陈晓平 , 王皖陵 .Linux 下实施流媒体的编程实现 . 安徽工业大学学报,2005,22(3):293-297.
[23] Alan Cox.Video4Linux Programming.alan@redhat.com.2000.
[24] 陈俊宏.Video Streaming探讨.http://www.jollen.org.2006.
[25] Murphy R R.Sensor and information fusion for improve division-based vehicle guidance. IEEE Intelligent Systems, 1998, Nov/Dec, 49-56.
[26] Guiducci A.Parametric model of the perspective projection of a road with applications to lane keeping and 3D road reconstruction.Computer Vision and Image Under standing, 1999,73(3):414-427.
[27] Giachetti A, Campani M.The use of optical flow for road navigation. IEEE Transactions on Robotics and Automation, 1998,14(1):34-48.
[28] 刘烽,吴昌英,毕文斌,许家栋.基于视觉识别的机动车辆智能驾驶系统.计算机工程与应用,2004,12:211-214.
[29] Turk M A, M orgenhaler D G. VITS-A vision system for autonomous land vehicle navigation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 1988, 10(3):342-361.
[30] 郝宝青.智能车辆视觉导航中道路与行人检测技术的研究.硕士毕业论文, 2006,6:18.
[31] 谢毅.基于 DSP 及机器视觉的道路识别与障碍物检测.硕士毕业论文.2007, 5:30-35.
[32] 苏开娜,丰丽军,刘忠云.公路汽车视觉导航中运动目标的检测与跟踪方法的研究.计算机工程,2001, 27(4):46-48,81.
[33] 黄席樾,柴毅,周欣,汪先矩,黄翰敏.汽车智能辅助操作系统中的障碍物检测.重庆大学学报(自然科学版),2000,23(4):123-126,137.
[34] 顾柏园.基于单目视觉的安全车距预警系统.毕业论文.2006:52-53.
[35] 刘燕,刘浩学.基于计算机视觉的单目摄影纵向车距测量系统研究. 公路交通 科技,2004,21(9):103-106.
[36] 由车祸谈及刹车.http://bbs.jdcjsr.com/showtopic.aspx?forumid=80&topicid= 2391&go=next.
[37] 辛德胜.驾驶员反应时间对行车安全的影响及检测系统.公路交通科技, 1999,(1):66~67.
[38] 华恒科技.嵌入式系统实验教程(ARM9+Linux 分册).合肥华开源恒信息技术有限公司,2004:30.
[2] Broggi A, Bertozzi M, Fascioli A, ConteG Automatic Vehicle Guidance. The Experience of the ARGO Automous Vehicle.World Scientific, Singapore, 1999.
[3] 胡永裕.怎样消除高速公路危险源.现代职业安全,2005,(7):56-57.
[4] 鲍吉龙,应延治等.基于 DSP 技术的汽车防撞雷达.微计算机信息, 2006,22(4-2):188-190.
[5] C.Thorpe.Vision and navigation: The Carnegie Mellon Navlab. Kluwer Acade- mic Publishers, 1990:83-93.
[6] L.Andreone, A.Broggi.Vehicle Detection and Localization in Infra-RedImages. The IEEE 5th International Conference on Intelligent Transportation System, Singapore, September 2002:3-6.
[7] 徐友春 . 智能车辆视觉与 GPS 综合导航方法的研究 . 博士毕业论文.2001:1-10.
[8] 杜春雷.ARM 体系结构与编程.清华大学出版社,2003.2:1-2.
[9] Peter Hank,Egon J?hnk. SJA1000 Stand-alone CAN controller. Application Note AN97076.Philips Semiconductors, December 1997.
[10] 朱华旭.CAN/LIN 总线技术在车内网络的应用研究.武汉理工大学学报, 2005,5:210-212.
[11] 饶运涛,邹继军,郑勇芸.现场总线 CAN 原理与应用技术.北京航空航天大学出版社, 2003.6:14-19.
[12] User’s Manual S3C2410X 32-bit RISC Microprocessor Revision 1.2,Samsung Electronic,2003.
[13] 求是科技.单片机典型模块设计实例导航.人民邮电出版社,2004.5:349-421.
[14] 王巍,秦贵和,范玮.基于 CAN 总线的汽车车身线束设计.制造业自动化,2005,27(7):71-74.
[15] Peter Hank, Egon J?hnk, SJA1000 Stand-alone CAN controller Application Note, Philips Semiconductors, 1997.
[16] Eisele, H. and J?hnk, E.PCA82C250/251 CAN Transceiver. Application Note AN96116. Philips Semiconductors, 1996.10.
[17] Data Sheet MCP2510, Microchip Technology, 1999.
[18] 潘巨龙,黄宁,姚伏天.ARM9 嵌入式 Linux 系统构建与应用.北京航空航天大学出版社,2006.8:12-13.
[19] AT89S52 Datasheets.ATMEL corporation.2001.
[20] 7805-LINEAR INTEGRATED CIRCUITS 3-TERMINAL VOLATGE REGUL–ATORS ETC.http://www.alldatasheet.com.
[21] VNH3SP30 - FULLY INTEGRATED H-BRIDGE MOTOR DRIVER. http:// www.alldatasheet.com.
[22] 陈晓平 , 王皖陵 .Linux 下实施流媒体的编程实现 . 安徽工业大学学报,2005,22(3):293-297.
[23] Alan Cox.Video4Linux Programming.alan@redhat.com.2000.
[24] 陈俊宏.Video Streaming探讨.http://www.jollen.org.2006.
[25] Murphy R R.Sensor and information fusion for improve division-based vehicle guidance. IEEE Intelligent Systems, 1998, Nov/Dec, 49-56.
[26] Guiducci A.Parametric model of the perspective projection of a road with applications to lane keeping and 3D road reconstruction.Computer Vision and Image Under standing, 1999,73(3):414-427.
[27] Giachetti A, Campani M.The use of optical flow for road navigation. IEEE Transactions on Robotics and Automation, 1998,14(1):34-48.
[28] 刘烽,吴昌英,毕文斌,许家栋.基于视觉识别的机动车辆智能驾驶系统.计算机工程与应用,2004,12:211-214.
[29] Turk M A, M orgenhaler D G. VITS-A vision system for autonomous land vehicle navigation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 1988, 10(3):342-361.
[30] 郝宝青.智能车辆视觉导航中道路与行人检测技术的研究.硕士毕业论文, 2006,6:18.
[31] 谢毅.基于 DSP 及机器视觉的道路识别与障碍物检测.硕士毕业论文.2007, 5:30-35.
[32] 苏开娜,丰丽军,刘忠云.公路汽车视觉导航中运动目标的检测与跟踪方法的研究.计算机工程,2001, 27(4):46-48,81.
[33] 黄席樾,柴毅,周欣,汪先矩,黄翰敏.汽车智能辅助操作系统中的障碍物检测.重庆大学学报(自然科学版),2000,23(4):123-126,137.
[34] 顾柏园.基于单目视觉的安全车距预警系统.毕业论文.2006:52-53.
[35] 刘燕,刘浩学.基于计算机视觉的单目摄影纵向车距测量系统研究. 公路交通 科技,2004,21(9):103-106.
[36] 由车祸谈及刹车.http://bbs.jdcjsr.com/showtopic.aspx?forumid=80&topicid= 2391&go=next.
[37] 辛德胜.驾驶员反应时间对行车安全的影响及检测系统.公路交通科技, 1999,(1):66~67.
[38] 华恒科技.嵌入式系统实验教程(ARM9+Linux 分册).合肥华开源恒信息技术有限公司,2004:30.