Enabling design-space exploration for robot SLAM - for accuracy, performance and energy

Prof Paul Kelly, Imperial College London

SLAM - simultaneous location and mapping - is a key platform for understanding 3D environments for a huge range of applications, spanning robotics and augmented reality and beyond. Building a really usable map of the environment requires "dense" methods currently feasible in realtime only on powerful hardware.  This talk will introduce SLAMBench, a publicly-available software framework which represents a starting point for quantitative, comparable and validatable experimental research to investigate trade-offs in performance, accuracy and energy consumption of a dense RGB-D SLAM system. SLAMBench provides a KinectFusion implementation in C++, OpenMP, OpenCL and CUDA, and harnesses the ICL-NUIM dataset of synthetic RGB-D sequences with trajectory and scene ground truth for reliable accuracy comparison of different implementation and algorithms. We present an analysis and breakdown of the constituent algorithmic elements of KinectFusion, and experimentally investigate their execution time on a variety of multicore and GPU-accelerated platforms. For a popular embedded platform, we also present an analysis of energy efficiency for different configuration alternatives.  This work is part of a larger research agenda aiming to push the limits of compiler technology up the "food chain", to explore higher-level algorithmic aspects of the design space and low-level implementation choices together, and I will present some preliminary results showing some of the potential of this idea.