Flying over the reality gap: From simulated to real indoor airships

Abstract

Because of their ability to naturally float in the air, indoor airships (often called blimps) constitute an appealing platform for research in aerial robotics. However, when confronted to long lasting experiments such as those involving learning or evolutionary techniques, blimps present the disadvantage that they cannot be linked to external power sources and tend to have little mechanical resistance due to their low weight budget. One solution to this problem is to use a realistic flight simulator, which can also significantly reduce experimental duration by running faster than real time. This requires an efficient physical dynamic modelling and parameter identification procedure, which are complicated to develop and usually rely on costly facilities such as wind tunnels. In this paper, we present a simple and efficient physics-based dynamic modelling of indoor airships including a pragmatic methodology for parameter identification without the need for complex or costly test facilities. Our approach is tested with an existing blimp in a vision-based navigation task. Neuronal controllers are evolved in simulation to map visual input into motor commands in order to steer the flying robot forward as fast as possible while avoiding collisions. After evolution, the best individuals are successfully transferred to the physical blimp, which experimentally demonstrates the efficiency of the proposed approach.

Learning Non-stationary System Dynamics Online Using Gaussian Processes

“…To obtain a realistic movement of the system a normal distributed noise term was added to the control commands to simulate outer influence like gust of wind. The system dynamics of the blimp were derived based on standard physical aeronautic principles [13] and the parameters were optimized based on a series of trajectories flown with the real blimp. To evaluate the predictive accuracy, we used 500 randomly sampled points and determined the mean squared error (RMS) of the predictive mean of the corresponding GP model relative to the ground truth prediction calculated by the simulator.…”

Velocity Controller for a Class of Vehicles

“…The equations (1) and (2) can be applied for ocean or marine vehicles as well as for airships during indoor experiments. Such examples can be found: (1) for underwater vehicles in [12,17], (2) for surface vessels in [7,18] or hovercrafts in [13,21], (3) for indoor airships in [23,31]. The control strategy is based on transformed equations of motion with the identity inertia matrix.…”

Underwater Robot Swarms: Challenges and Opportunities

“…This is why airships are usually very sensitive to wind and cannot travel efficiently at high speeds. However, because of their ability to naturally float in the air, small airships (also called blimps when there is no rigid structure inside the hull) have been widely used as research platforms in aerial robotics (Bermudez i Badia et al, 2005;Iida, 2003;Van Der Zwaan et al, 2002;Zhang and Ostrowski, 1998;Zufferey et al, 2006a). In particular, Melhuish and Welsby (2002) have been using small indoor blimps to demonstrate flocking behaviours (Fig.…”