On the Application of Efficient Neural Mapping to Real-Time Indoor Localisation for Unmanned Ground Vehicles

Abstract-- Global localisation from visual data is a challenging problem applicable to many robotics domains. Prior works have shown that neural networks can be trained to map images of an environment to absolute camera pose within that environment, learning an implicit neural mapping in the process. In this work we evaluate the applicability of such an approach to real-world robotics scenarios, demonstrating that by constraining the problem to 2-dimensions and significantly increasing the quantity of training data, a compact model capable of real-time inference on embedded platforms can be used to achieve localisation accuracy of several centimetres. We deploy our trained model onboard a UGV platform, demonstrating its effectiveness in a waypoint navigation task, wherein it is able to localise with a mean accuracy of 9cm at a rate of 6fps running on the UGV's onboard CPU, 35fps on an embedded GPU, or 220fps on a desktop GPU. Solutions that involve the placement of fixed retrieved images to refine the final estimate [14] [15] [16]; markers or beacons, such as ultra-wideband positioning [1], ultrasonic tracking beacons [2] or visual markers [3] can 5. 2D - 2D Implicit Map Localisation, that we refer to in this facilitate accuracy ranging from centimetres to metres, and work as neural mapping, estimates pose via a neural require specialist hardware be placed within the environment network that has learned an implicit representation of a and in some cases on agents themselves.