IO-VNBD: Inertial and Odometry benchmark dataset for ground vehicle positioning

Low-cost inertial navigation sensors (INS) can be exploited for a reliable tracking solution autonomous vehicles. However, position errors grow exponentially due to noises in the measurements. Several deep learning techniques have been investigated mitigate better [1-10]. these studies involved use of different datasets not made publicly available. The lack robust benchmark dataset has thus hindered advancement research, comparison and adoption vehicle positioning based on navigation. In order facilitate benchmarking, fast development evaluation algorithms, we therefore present first its kind large-scale information-rich odometry focused public called IO-VNBD (Inertial Odometry Vehicle Navigation Benchmark Dataset).The was recorded using research equipped with ego-motion roads United Kingdom, Nigeria, France. include GPS receiver, sensors, wheel-speed amongst other found car as well receiver an android smart phone sampling at 10HZ. A diverse number scenarios dynamics are captured such traffic, round-abouts, hard-braking etc. road types (country roads, motorways etc.) varying driving patterns. consists total time about 40 hours over 1,300km extracted data 58 4,400 km smartphone data. We hope that this will prove valuable furthering correlation between displacement related