Neuromorphic computing is a promising solution for reducing the size, weightand power of mobile embedded systems. In this paper, we introduce a realizationof such a system by creating the first closed-loop battery-poweredcommunication system between an IBM TrueNorth NS1e and an autonomousAndroid-Based Robotics platform. Using this system, we constructed a dataset ofpath following behavior by manually driving the Android-Based robot along steepmountain trails and recording video frames from the camera mounted on the robotalong with the corresponding motor commands. We used this dataset to train adeep convolutional neural network implemented on the TrueNorth NS1e. The NS1e,which was mounted on the robot and powered by the robot's battery, resulted ina self-driving robot that could successfully traverse a steep mountain path inreal time. To our knowledge, this represents the first time the TrueNorth NS1eneuromorphic chip has been embedded on a mobile platform under closed-loopcontrol.
展开▼