Research On Active Vehicle Collision Avoidance System Based On Pulsed Laser Ranging

In recent years, vehicle was becoming more and more popular in transportation and travelling, Traffic accidents were also becoming one of the most serious killers of humans’ lives and property. To avoid traffic accident as much as possible, vehicle active collision avoidance system is gradually becoming a focused research in the field of driving safety. This paper aims to study automotive active collision avoidance system based on pulsed laser ranging. On basis of designing pulse laser rangefinder based on SPLJLL90₃ and AD500₉, this paper have done detailed research on collision avoidance mechanism, error correction and false alarm controlling.According to the principle of pulsed laser ranging, this paper designed a pulsed laser rangefinder whose accuracy is 0.0195m, conducted experiments for each module then showed the output waveform, and designed vehicle-steering measuring module and vehicle-speed measuring module to collect vehicle’s current driving state. Then the paper analyzed the effective range, accuracy and error source of rangefinder.On this basis, the paper analyzed the model of the car’s braking, deduced the alarm and braking safety distance, and determined a collision avoidance algorithm based on the relative velocity. Combined with car moving model, the paper studied several correction methods on distances and velocities, including Mathematical Morphology Curve Fitting, correction of distances and velocities based on Least-squares curve fitting and correction of distances and velocities based on kalman filter. These methods achieved good results.Finally, this paper proposed a kind of rangefinder’s Installation location called three-beams and angular-difference. On this basis, we analyzed several situations which would cause false alarm and designed corresponding false alarm controlling algorithms, including:ramp recognition, non-target recognition, turning-avoidance false alarm controlling, changing-lanes false alarm controlling. Through experiments, these algorithms achieved good results.