Abstract: Coal production is a typical highrisk and hard work. Existing coal mine equipment requires skilled workers to directly or indirectly participate，which has high security risks and costs. Considering safety and efficiency of the industry，using robots in coal production instead of workers is an inevitable way for coal mining companies to achieve “fewerpeople” or even “unmanned” production. It is also in line with the basic goal of “digital mines”. In the autonomous operation of coal mine robots，environmental perception technology，which would realize robots’ safety status assessment and provide basic data for decisionmaking and planning，is used to obtain workspace information. Moreover，path planning is the basis for robot movement and a key technology connecting environment perception and underlying control. The working principle，functional requirements，task objectives and environmental complexity of 5 types of coal mine robots（i.e. roadheader robots，shearer robots，transport robots，inspection robots and rescue robots) are summarized in this paper. We organize and analyse the important research of coal mine robots focusing on the environment perception and path planning in the past 15 years. Finally，the future development direction of key technologies of coal mine robots is prospected from 3 aspects：perception and cognition，planning and decision and system integration. The development trend of perception and cognition not only requires coal mine robots to have perception capabilities，but also requires robots to complete cognitive learning based on a small amount of available information outside of regular events. The key challenge is to develop multisensor fusion technology and fewshot learning methods. The development direction of planning and decision can be divided into singletarget planning and multitarget decisionmaking. Coal mine robots use perception and cognitive information to achieve pointtopoint path planning and obstacle avoidance，and complete multitask optimization according to work requirements. Thereby，the working efficiency of coal mine robots can be improved. The development of robot collaboration systems and communication interactive systems is the main challenge for system integration. Through a complete internal and external network system，realtime and closedloop interaction among coal mine managers，technical workers and robot equipment and cooperation between multinumber/multitype of coal mine robots will be completed.