3results about How to "Solving path planning problems" patented technology

The machining system comprises an intelligent conveying system, an intelligent threading system and an intelligent grinding system, the three systems achieve real-time linkage through sensing data interaction and algorithm collaborative scheduling, and full-process closed-loop control is formed. Wherein the intelligent conveying system is used for accurate conveying and specification identification of reinforcing steel bars and cooperative scheduling of multiple conveying units, and reinforcing steel bar blanks which are accurate in positioning and stable in posture are provided for subsequent threading and polishing procedures; the intelligent threading system is used for high-precision threading machining of the ends of the reinforcing steel bars, the state of the reinforcing steel bars is sensed in multiple dimensions, machining parameters are adjusted in a self-adaptive mode, threading defects are avoided, and the thread forming quality is guaranteed. The intelligent grinding system is used for precisely grinding the surfaces and the ends of the steel bars, it is ensured that the quality of the ground surfaces meets the engineering requirements by sensing the grinding contact state and dynamically optimizing grinding parameters, and the construction efficiency is improved through the machining system.

This invention belongs to the field of fire rescue technology, specifically providing a method and system for fire emergency rescue route selection based on deep reinforcement learning. The method includes: abstracting real road environment information into network information as the environment input for deep reinforcement learning; considering the shortest time to get fire rescue vehicles to the fire rescue point as the optimization objective; establishing road evaluation indicators based on information such as the length of each route, traffic flow, number of lanes, and average vehicle speed, and designing a reward function based on these indicators; modeling the optimization objective as a Markov decision process, and using a PPO model to solve for the optimal strategy. This invention comprehensively considers the information differences between various routes and takes into account real-time road conditions, enabling rapid completion of fire rescue route planning.

This invention discloses a control method and a vibratory feeding device for a vibratory feeding apparatus, relating to the field of automatic feeding technology. The control method for the vibratory feeding apparatus includes the following steps: after calibrating multiple cameras and multiple adsorption units, controlling the vibratory feeding apparatus to enter a working mode; acquiring material state images captured by multiple cameras and identifying target materials in the multiple material state images; planning a picking path based on the pixel coordinates of multiple target materials; converting the pixel coordinates of multiple target materials into physical coordinates and matching multiple adsorption units with multiple target materials; controlling the picking device to move according to the picking path to complete the picking of multiple target materials. This individual calibration method improves the picking accuracy and solves the problems of positioning and allocating materials under visual guidance by multiple cameras and multiple adsorption units, as well as the path planning problem. Path planning can be performed at the pixel coordinate stage, optimizing the length of the picking path and thus shortening the picking cycle.