210 likes | 219 Views
This paper presents the implementation of an in-vehicle multi-sensor information fusion gateway for cooperative driving, enhancing active safety and reducing collision risks.
E N D
Implementation of In-Vehicle Multi-sensor Information Fusion Gateway for Cooperative Driving Authors: Ting-Ying Wei, Zhi-Liang Qiu, Chung-Ping Young Presenter: Zhi-Liang Qiu
Outline • Introduction • Multi-Sensor Information Fusion • Hardware architecture • Vehicle Gateway • Electronic Control Units (ECUs) • Communication • Sensor • Software • Conclusion
Introduction • The traffic crashes caused by the human factor in US are about 80%-90% • The safety concern is the most important issue for vehicle operation • Cooperative driving • Enhance the driving information • Increase the response time
Introduction • Precaution warning or compensatory control before the pre-crash stage of vehicle safely • Heterogeneous multi-core processor: • One ARM core for I/O control and system management • Two DSP cores for intensive computation of information fusion • Signal processing • Location mapping • Trajectory prediction • Risk assessment
Introduction • Two types of tasks: • Real-time • Non-real-time • The ARM processor: • General Purpose Processor (GPP) for the non-real-time tasks • PAC DSP: • Special Purpose Processor(SPP) for the real-time tasks • Electronic Control Unit(ECU): • Gathering the surrounding information
Multi-Sensor Information Fusion • Vehicle gateway implements the information fusion • In-vehicle sensor data • Wireless communication packets • GPP will inform the data to Information fusion engine in SPP whenever the data is received • SPP generates several safety indexes and produce warning messages
Multi-Sensor Information Fusion • Procedure: • Retrieving information • Integrating the information in a time sequential linked list • Cross-evaluating the information • Mapping each car in the tree structure onto a highway map • Predicting the upcoming path • Generating warning messages • Broadcasting to surrounding vehicles
Multi-Sensor Information Fusion • Procedure:
Multi-Sensor Information Fusion • Time synchronization base on the GPS signal from each car • Image processing DSPs detect the lane departure and blind spot • Camera frame rate: 30fps • Sensor sampling rate: 200Hz
Multi-Sensor Information Fusion • Location mapping: • GPS data is only be used for global positioning route planning or track recording • Merging the signals from radar, accelerometer, wheel speed sensor, and yaw sensor generates the relative distance, position and velocity
Multi-Sensor Information Fusion • Trajectory prediction: • Using the current and past vehicle information as well as the circumstance event to estimate the next few seconds path • If there are any specific events, like emergency brake, we will receive the warning from the event car
Multi-Sensor Information Fusion • Risk assessment:
Vehicle Gateway • The vehicle gateway is a core of this system, it includes many different interfaces of communication as shown below
Vehicle Gateway • The realization of vehicle gateway
Vehicle Gateway • PAC Duo DSP development platform • The PAC Duo platform is developed by Industry Technology Research Institute (ITRI) in Taiwan • Each DSP core communicates with ARM core through AXI-AHB bus
Electronic Control Units • Microchip PIC18Fxx8 microcontroller • Renesas R32C/118 The microcontrollers are the middle layers between PAC Duo and the surrounding sensors
Communication • There are several wired or wireless of communication interfaces are required • CAN serial bus protocol • OBDII • IEEE 802.15.4 • DSRC • 3G
Sensors • Different types of sensors are used for acquiring vehicle diagnosis or driving status • Infrared night vision camera • The GPS receiver
Software • The software is separated into several parts • Non-real-time OS, Meego • Open source Linux-based mobile OS • Supported by Intel and Nokia • Provide the GUI • Real-time on PAC DSP • Non-OS on each microcontroller board
Conclusion • A heterogeneous multi-core platform is establishing. • The multi-sensor information fusion is implemented for computing the relative position, velocity and acceleration to the host vehicle. • Assess the potential of collision risk and provide the cooperative driving feature to enhance the active safety