Team Rivoli

 

Inertial & Vision Research for Indoor Localization

The RIVOLI project (Inertial & Vision Research for Indoor Localization) proposed by iXblue and ONERA is aimed at developing a pedestrian localization system for non-collaborative indoor / outdoor environment and test it during the 3 competitions of the MALIN Challenge.

RIVOLI is based on the recognized iXblue know-how in navigation systems and ONERA know-how in embedded vision algorithms. The hardware solution is based on these two main components:

-an iXblue compact FOG50 inertial navigation system (FOG = Fiber Optic Gyroscope), having a high performance (0.1°/h) within a small volume

-an ONERA 3D vision system measuring displacements and building up environment map

The first objective is to design progressively a smart hybridization between inertial and vision data through the successive competitions, in order to maintain high position accuracy. The intrinsic high INS precision can maintain localization accuracy during outages of the vision measurements (in case of bad visibility or indoor-outdor transitions).

Around the core inertial-vision solution, additional sensors will be evaluated in order to improve robustness and performance of the RIVOLI system: additional vision sensors, podometers, Doppler velocity sensors, etc… Selecting the right additional sensors for the Challenge will be done after a multi-criterion analysis looking at cost, volume, consumption, hardware and firmware integration facility, and contribution to the localization accuracy.

The consortium proposes a first INS / vision efficient solution able tomeet the Challenge objectives. Nevertheless, the objective is to improve this solution in order to maintain requested accuracy while scenarios difficulties increase. For each competition, an upgraded system will be integrated, including the recent research advances:use of already mapped environment for position update in the second competition and a global fusion with all additional sensors in the last one.

The RIVOLI project takes place in an operational context of a civilian or military pedestrian moving in a non-cooperative or even hostile environment with a need for a metric localization precision and an environment mapping, easy-to-use and as cheap as possible. Operational constraints, ergonomy and adequate INS performance/cost ratio will be key drivers during the whole project in order to reach at end of project the most efficient solutions.

The RIVOLI project proposed by iXblue and ONERA satisfies all the requirements of the MALIN Challenge by mixing innovation and experiments. At the end of the project, both partners’ developments will benefit from technological improvements that can lead to future military or civilian applications dealing with 3D mapping and localization of a pedestrian in an unknown environment.