Abstract: For the technical problem of low intelligence in the process of drilling and pressure relief in high stress mines, the research status of pressure relief technology and equipment at home and abroad is summarized and analyzed in this study. It is pointed out that the development of high-performance, highly reliable, and efficient fully autonomous drilling system of anti-impact drilling robot is an important development direction to solve the problem of rock burst prevention and control. To this end, the five key technologies that affect the performance of the drilling system, namely “the precise recognition of hole position, the precise perception of drilling tool posture, the wireless electromagnetic intelligent detection, the intelligent recognition of drilling tool operation status, and the precise control of the drilling system” have been summarized, and the solutions and methods have been provided. In response to the problem of accurate identification of pressure relief holes in complex and harsh environments, a SinGAN model for pressure relief hole image sample expansion is developed, which integrates image size adjustment and multi-stage training modes. The Faster RCNN optimized by multi-layer feature fusion is introduced, and a hole position recognition model based on an improved SqueezeNet lightweight network architecture is constructed to achieve an accurate and fast recognition of pressure relief hole positions. To address the issue of precise perception of drilling tool posture, the unscented Kalman filter optimized by improved gradient descent algorithm is designed for the initial alignment of Inertial Measurement Unit (IMU). Multiple IMU spatial array layouts are designed, and a BP neural network-based compensation method for drilling tool posture error is studied, aiming to improve the accuracy of drilling tool posture calculation and achieve a precise drilling pressure relief. Aiming at the precise detection of drilling conditions in complex geological environments, a wireless elec-tromagnetic transmission system architecture for underground measurement while drilling in coal mines has been established. The principles of adaptive modulation of weak electromagnetic wave signals and high-speed bidirec-tional electromagnetic transmission technology have been explored, and the measurement principles and imple-mentation processes of geological parameters, geological parameters, and engineering parameters at the bottom of the hole have been investigated. To address the issue of identifying the operational status of drilling systems, a multi-domain feature extraction architecture for drilling signals in the time domain, frequency domain, and time frequency domain, as well as a deep network advanced feature extraction architecture, have been constructed. In addition, the key component health status assessment and fault diagnosis techniques for drilling systems have been proposed, and a prediction model for sticking risk factors based on an improved bat optimized long-term and short-term memory network has been built to accurately predict the stuck status of pressure relief drilling tools. In terms of the issue of precise control of drilling systems, the working principle of the hydraulic system of the drilling system is analyzed, and a precise control scheme for the drilling system considering the characteristics of coal and rock is formulated. The principle of solving the optimal control parameters of the drilling system based on torque and position is explored, aiming to achieve intelligent collaborative control and parallel operation of the drilling return system and feed system.