Abstract: Car driver's motion estimation is important for developing safety driving systems .We propose to use particle filters for this purpose .Several sensors, such as camera to capture driver'sface, velocity sensor using Doppler Effect of microwave for driver's foot motions, and microphones to record and to identify drivelsvoice, are placed at experimental cabin of drivingsimulator with three display panels. From thesesensor signals, particle filters estimate thein formation to be obtained based on someel aborated state space models. So obtainedinformation are gathered to server machine for making global recognition of the driver's body motion. We have implemented the data gatheringsystem based on server-client framework onTCP/IP.Keywords: Safety Driving, Support System, Body Motion, Particle Filter. I Introduction
Assistance of safety driving by relaxing card river's workload to prevent traffic accidents is worth not only for the car driver him/herself but also other traffics including other cars, bikes, bicycles and pedestrians. Recent development of technologies arrows us to use various functional sensors, such as camera, laser range finder, and soon, with reasonable price and easy way to use, and computers have become small and high performance. Thus we can use such functionalsensors as well as highly sophisticated signal processing methods for supporting safety driving. For example, detection system of for wardobstacles by measuring forwarding cars with camera and/or laser range finder, parking assistant system by merging multiple camera images of surrounding of a car into a single image of topview, detection of eye motion and/or eyelid fromcar driver's face image to observe the driver's condition, and so on. Some of these examples have already been in consumer products while some others have been in a phase of research and development.
It Seems that in these conventional systems for safety driving rather simple signal processing methods have been employed due to strictly shorter sponse time and limitation of ECU (Electronic Control Unit) performance in time. Most of actual signals such as camera image and laser range measurement include various aspects of noises, sowe need to use some method to reduce such noises .
……
展开