Research On Obstacle Avoidance And Fire Early Warning Of Autonom Ous Patrol Robot

The warehouse is the place where the material is stored centrally,and it is usually confined space,and most of them are combustible.Once the fire happens,it will cause huge loss.At present,there are many problems in the warehouse fire warning,such as low level of automation,the monitoring blind area,high false alarm rate and false alarm rate.In order to find out the fire danger in time,ensure the safety of the goods and personnel in the warehouse,it is of great significance to study the fire early warning robot with autonomous patrol function.In this thesis,the wheeled patrol robot is designed,which can avoid the obstacles in the road,and make an early warning of the possible fire in the warehouse.The obstacle avoidance is the precondition of completing the fire early warning,In this thesis,according to the characteristics of the warehouse environment,the obstacle avoidance scheme is designed: Ultrasonic sensors are used to detect obstacle,then detect the location of obstacle by visual method.First,camera capture obstacle image,the edge of the obstacle is extracted by Sobel edge detection algorithm.On this basis,the edge linear feature of the obstacle is obtained by using the Hough algorithm,and find feature point;Secondly,the pinhole model of monocular vision is established to determine the location of obstacle;Finally,according to the position of the obstacle avoidance strategy,fuzzy obstacle strategy control motor to achieve obstacle avoidance.In this thesis,patrol robot fire linkage warning strategy is designed,real time detection temperature,smoke,and humidity in the process of robot movement,if fire is detected,patrol robot stops moving,and send out fire linkage control command,opening the relevant fire linkage equipment.Research on flame,smoke detection and recognition algorithm based on image,the characteristic model of flame and smoke is established,help to improve the warehouse fire automatic warning ability.In this thesis,the FPGA hardware integration of patrol robot obstacle avoidance and fire warning function is completed,Verilog HDL language programming algorithm,ModelSim software and SignalTap II logic analyzer debugging each function module.The overall effect of the patrol robot system is verified under the experimental condition,patrol robot can accurately avoid obstacle,when fire is detected,robot sends out fire linkage control command.Through flame and smoke image detection experiment,the visual detection and recognition of fire danger is realized.