| The rapid development of robots brings about revolutionary changes in industry as well as daily lives.Sweeping robots are gradually taking over traditional manual cleaning works.Due to the continuous reduction of cost and improvement of working efficiency,sweeping robots are increasingly accepted by consumers and have a broad market.In recent years,sweeping robots designed abroad are capable of positioning and planning.Also,they have a remarkable cleaning efficiency,however,as well as the price.In contrast,the cost of domestic sweeping robots is acceptable,but their performance is far from ideal.In this thesis,a hardware circuit for sweeping robot with visual positioning and visual obstacle avoidance is designed.It not only provides an algorithm verification platform but also implements a mobile remote surveillance system with the camera.The main work of this thesis is as follows:(1)The main functions of the verification platform include:collision,drop,off the ground and other emergency state awareness,movement mileage and course information acquisition,visual positioning and obstacle avoidance and other visual information acquisition,movement execution,floor cleaning interface,automatic recharge,system power distribution and monitoring systems.According to these functions,corresponding sensors will be selected and driving codes will be designed.(2)The ARM+STM32 dual-core architecture mode is used to process data,there are mainly two kinds of data:image information of large amount and non-image information of small amount.STM32 core(model:STM32F103VET6)is applied to realize the driving and data acquisition of many sensors other than the image,as well as to control the movement of the wheel motor;ARM core(model:S5PV210)is applied to achieve the camera image acquisition and access to a wireless network in the surveillance system.These two cores are connected by serial port:ARM core receives sensor data,image positioning,and obstacle avoidance information from the STM32 part;and after these data are comprehensively processed.a motion decision is generated and sent to the STM32 to perform forward,backward,turning,etc.of the sweeping robot.(3)The surveillance system is implemented with the Qt architecture.The sweeping robot is designed as a server,while Windows and Android as the clients;the Qt architecture deployment on the server is completed by cross-compiling Qt source packages.The real-time transmission of video is based on RTP/RTCP protocol;the two-way PWM drive in embedded Linux server is designed to realize horizontal and pitch angle control of the servo of the PTZ-mounted camera.The client uses a hybrid Gaussian model to extract the target foreground combined with the dynamic interframe difference method,detecting whether the monitoring area changes,and only real-time recording of valid dynamic pictures are performed;the state information of the sweeper is continuously received and displayed using the TCP protocol.With specific sensors,the hardware circuit together with the software driver can realize the environment perception of the sweeping robot,obtaining visual information,mileage,heading angle,and emergency status.It provides an effective verification platform for the monocular visual positioning and monocular visual obstacle avoidance algorithm.And a wireless network surveillance system is designed to achieve real-time control of the sweeping robot.According to the test:the video frame rate is an average of 20 frames/s and the display is smooth;dynamic target real-time detection takes about 43ms;monitor field of horizontal,pitch with ±90°. |
PDF Full Text Request