Pseudo Real-time Visual Estimation Of Mobile Robot Under Network Delay

Due to the incomparable advantages over human, mobile robot technology hasbeen applied to a great many fields. Revolution on robot technology never stops.However, the bottleneck of technology and the non-equalizing problem go side byside in every era. Fortunately, smart man always found ways to solve such problem inparticular situations. One of the problem is how to improve the visual real-timefeature from a remote mobile robot system.This article is about to solve the problem in a new way based on visual principles.In order to improve the human operator’s experience.In the system, the local host part is played by a Linux OS computer with Internetconnection. A mobile robot with visual sensor played the remote one. The local hostreceive a UDP package containing image information and a timestamp whichrecorded on capturing the image. Comparing the timestamp received with local timesynchronized before, the local host system knows the specific delay between currenttime and image recorded time. Using such delay info and the speed of the mobilerobot, local host start to estimate the pseudo real-time image by processing the image,mostly, zooming and relocating operations.In such ways the system give the human operator an acceptable low delayexperience instead of a non-meaningful high delayed video to reference.The local host computer have a Qt programed console with the ability ofconnecting and controlling mobile robot by internet, receiving image and displayingthem in order over the console.The remote part equipped with a ARM development board work as upper layer.Include serving the Internet connection, controlling command transfer and work likevideo capturing and coding. The lower layer is played by80C51boards.Using CANnetwork designed by author, equipped with a speed control module. Due to the usageof CAN network, expandability in future experiment is guaranteed.Completing all parts of work above, the testing results show that the system workwell on estimating. Solving the delayed visual problem in a certain degree.