Automatic parking method based on fusion of vision and ultrasonic perception

[0069] Embodiment: The automatic parking method of this embodiment based on the fusion of vision and ultrasonic perception can complete the parking space search in a specific scene and complete the automatic parking operation according to the parking space selected by the user, based on the system hardware architecture Such as figure 1 As shown, it mainly includes a 360° surround view system, an ultrasonic radar system, an automatic parking controller, a vision coprocessor, and an actuator. Among them, the surround view system contains a total of 4 fisheye cameras, each camera has a horizontal field of view angle of 180°, which is transmitted to the vision coprocessor (which can be integrated with the automatic parking controller) through the HD-CVI protocol (or LVDS) After processing, the relevant information (including visual parking space, target parking space and driving area) is sent to the automatic parking controller. The ultrasonic radar system includes a total of 12 ultrasonic radar sensors. Each sensor has a field of view of 60° and a maximum detection distance of 5 meters. The signal is transmitted from the CAN bus to the automatic parking controller and merged with the visual results to be generated by the trajectory planning module. Parking track. The vehicle control signals output by the automatic parking controller are respectively sent to the corresponding actuators through the motor controller (MCU), the vehicle controller (VCU), the brake system controller (ESC) and the steering system controller (EPS).

[0070] Automatic parking method based on the fusion of vision and ultrasonic perception, such as figure 2 As shown, the input is the surround view system scene image, the ultrasonic radar system signal and the wheel speed and steering wheel angle sensors, and the output is the target vehicle speed, target acceleration, target steering wheel angle and gear control signal, which specifically includes the following steps: