| The intelligence of coal mine equipment is one of the important directions to promote the development of smart mine,which is important to improve the level of coal mine safety production and realize the less personalized and unmanned management of coal mine.The coal mine drilling robot is a key piece of equipment for underground gas control.Achieving its intelligent and fully autonomous workflow can significantly reduce the safety risks of gas extraction and improve efficiency.The autonomous navigation system is the core module of the intelligent upgrade for the drilling robot,but there are no proven software and hardware design solutions and related technical methods.In the complex and harsh underground environment,advanced autonomous driving technology and related solutions on the ground are difficult to directly penetrate into the underground application.Existing autonomous navigation solutions in underground shafts are only designed and implemented for wider roadwyas or wheeled autonomous mobile equipment such as unmanned scrapers,and no research has been carried out on narrower working faces and large aspect ratio tracked differential mobile robots.Therefore,this thesis addresses the structural characteristics of the drilling robot and its special requirements in autonomous walking,focusing on three key technical issues: motion planning,autonomous obstacle avoidance and visual tracking.The main contents are as follows:For the actual requirements of autonomous navigation of coal mine drilling robots,its working environment,motion characteristics and working condition requirements are analysed.The modular functional architecture of the autonomous navigation system is designed according to the actual working conditions.The application scheme of mapping and positioning module,the research scheme of motion planning module and autonomous obstacle avoidance module are compared,analyzed,and selected.Design and development the hardware and software architecture for the autonomous navigation system of the drilling robot.Developed a functional verification mobile test platform for an autonomous navigation system with the characteristics of motion control of drilling robots.The programme has been developed and the conditions created for carrying out research into key technologies for autonomous navigation.Aiming at the motion planning problem of drilling robots in the narrow and confined space of coal mines,a smooth motion planning method based on Hybrid A*global planning combined with MPC trajectory tracking control is studied,and a smooth motion planning system that meets the actual motion constraints of drilling robots is established.To avoid the robot from turning in place during the movement,a Hybrid A* graph search combined with collision detection,heuristic convergence,and gradient optimization smoothing path planning method is proposed to implement global path planning for the robot.To ensure that the robot can follow the target path,an optimisation problem for a non-linear MPC trajectory tracker in the discrete time domain is derived and the associated solution scheme is elaborated.The effectiveness of the smooth motion planning system designed in this thesis is verified through simulation and ground site tests,and the study can meet the functional requirements for continuous fixed-point parking of the drilling robot.Aiming at the problem of autonomous obstacle avoidance by drilling robots in the unstructured environment of coal mines,an offline obstacle avoidance algorithm based on three-dimensional point cloud information is studied,and an autonomous obstacle avoidance system for the complex terrain of drilling robots is established.To ensure that the robot can fully perceive the terrain information and generate obstacle avoidance paths,a maximum arrival probability local planning method based on multi-line laser point cloud elevation information map is studied and proposed.An offline obstacle avoidance method based on an offline trajectory library of Bézier curves and a trajectory indexing rule based on multiple combination constraints are proposed to achieve smooth robot motion control and to keep away from obstacles on both sides.The effectiveness of the autonomous obstacle avoidance system proposed in this thesis is verified through simulations and field tests,and the study can satisfy the autonomous navigation capability of mobile robots with differential motion models in unstructured terrain.Aiming at the problem of collaborative control of the front and rear vehicles of split drilling robots,a visual object tracking method for low-illumination environment in underground is studied and proposed,and the tracking control system of front and rear vehicles of drilling robots is established.To tackle the influence of underground low-illumination environment on the visual tracking algorithm,a low-illumination long-term correlation filter tracker LLCT is proposed,and a object tracker based on translation-scale-long-time filter combined with low-illumination image enhancement module is designed.To obtain high real-time image processing effects,a fast image enhancement method with target brightness detection strategy is proposed.A front and rear vehicle vision tracking system based on April Tag 2D target to obtain depth information and combined with fuzzy PID for distance control is designed.The performance of LLCT is verified in open datasets and self-built low-illumination image sequences,and the effectiveness of the tracking control system proposed in this thesis for target object following control is verified in night scenes.To verify the practicality and feasibility of the algorithms and related systems proposed in this thesis,functional verification tests and performance tests of the autonomous navigation system on a mobile test platform and a live drilling robot are launched.The test and application test are carried out in the simulated roadway environment and underground field environment of coal mine,and the results show that the motion error of the smooth motion planning system proposed in this thesis can be controlled within 15 cm during continuous fixed-point parking.The autonomous obstacle avoidance system enables the drilling robot to turn collision-free in right-angled roadways.The autonomous navigation system proposed in this thesis could guarantee an overall positioning accuracy of ±20 cm for the entire travel-drilling process in the interconnection tests of the drilling robot system.The front and rear vehicle visual tracking control system can realize the safe following and stopping of the rear vehicle of the drilling robot within the set distance.The final underground field application tests have verified that the proposed autonomous navigation system is capable of path planning,spot parking and autonomous obstacle avoidance in the underground coal mine environment and meets the application requirements.In particular,the proposed LLCT method can successfully follow a target in a low illumination environment downhole and obtain a target movement trajectory.Through the research on the software and hardware architecture of the autonomous navigation system,the smooth motion planning method in the narrow and restricted space of the underground,the autonomous obstacle avoidance method in the unstructured underground terrain and the visual tracking control system in the low-illumination environment of the underground,the autonomous navigation system of the drilling robot applicable to the complex restricted environment of the underground coal mine was designed and developed,and the autonomous walking function of the drilling robot is realised.The relevant theoretical and technical results obtained in this thesis are of great significance in promoting the development of intelligent and robotised large mobile equipment downhole.The thesis has 150 pictures,34 tables,and 256 references. |
PDF Full Text Request