In complex mine environments， the autonomous perception and navigation system of mining cars is a key technology to ensure their efficient， accurate， and safe operation. This system enables mining trucks to autonomously adapt to the special conditions of the mine， significantly improving production efficiency while reducing operational risks. This paper puts forward a set of automatic sensing and navigation system for tracked mine cars， which is suitable for complex mines， including the sensing module， edge computing unit， communication module and back end data processing part. Real time perception and navigation of mining vehicles are achieved through the “edge+cloud” collaborative computing strategy. The test results indicate that the designed autonomous perception and navigation system can meet the automatic navigation task requirements of tracked mining cars in industrial and mining environments.