Multi-Modal Real-Time Collision Avoidance System in Working Space Based on Multi-Sensor Fusion

In order to improve the performance of the anti-collision system and reduce casualties and equipment damage caused by collision accidents, a multimodal real-time anti-collision system in the working space based on multi-sensor fusion is studied. In terms of hardware, laser radar, ultrasonic area ranging sensor and visual auxiliary monitoring integrated equipment are used to provide all-round visual assistance. In terms of software, the Kalman filter algorithm is used to fuse multimodal sensor data to obtain multidimensional information of obstacles. Based on multidimensional information, the dynamic danger distance is calculated, combined with the braking distance of the excavator, the danger distance is compared, and an anti-collision warning mechanism is established to ensure the anti-collision of the excavator. The test results show that the multimodal real-time anti-collision system in the working space based on multi-sensor fusion has a faster anti-collision warning response speed, a high detection range, and a good visual anti-collision effect, which meets the safety needs of actual operations.