Abstract: In order to meet the survey needs of passenger vehicles chassis at the accident scene, an autonomous survey robot for images acquistion of passenger vehicle chassis was developed. The image information of vehicle bottom was collected by a pan-tilt platfrom semi-submersible two degree of freedom, and the relative position between the robot and the chassis was estimated based on the motion coordinates of the robot and the occlusion area information of the infrared sensor array. According to the position information between the infrared sensors, a three-step calibration method of robot relative to vehicle attitude based on the sensor trigger timing was adopted to solve the drift problem of the inertial navigation module under the iron car body. The multi-sensor combined information was used to realize the robot’s compound perception of the environmental obstacle information, and the fuzzy control idea was used to realize the robot’s autonomous obstacle avoidance operation. The autonomous running mode test of the developed robot was carried out on asphalt pavement and concrete pavement. The results show that the developed robot can automatically realize the survey and traversal of the chassis of passenger vehicles, and the high-definition chassis image information collected can meet the requirements of traffic police department for accident investigation and evidence collection.