In this paper, the authors present the performance analysis of a Vertically Offset Overlapped Propulsion System (VOOPS)-based quadrotor in an aerial mapping mission. The dynamic model of the VOOPS quadrotor with the effect of overlapping propellers and the profile drag has been derived and simulated. A path-tracking mission is taken as an example for aerial survey. The controller used for this task is presented, followed by the response study of the attitude and the position controller with standard test inputs. A graphical interface has been built to select the area to be mapped by defining a polygon around it, and waypoints for lawnmower type survey grid were generated based on the direction of wind. The path-tracking algorithm is presented along with course correction and simulations were performed with both conventional and VOOPS quadrotor. An experimental vehicle based on the proposed VOOPS concept has been built, tested on the same path, and the results are discussed. The results show that the VOOPS quadrotor is capable of performing the aerial mapping mission with quick response and good accuracy.
Reinforcement learning algorithms are sensitive to hyper-parameters and require tuning and tweaking for specific environments for improving performance. Ensembles of reinforcement learning models on the other hand are known to be much more robust and stable. However, training multiple models independently on an environment suffers from high sample complexity. We present here a methodology to create multiple models from a single training instance that can be used in an ensemble through directed perturbation of the model parameters at regular intervals. This allows training a single model that converges to several local minima during the optimization process as a result of the perturbation. By saving the model parameters at each such instance, we obtain multiple policies during training that are ensembled during evaluation. We evaluate our approach on challenging discrete and continuous control tasks and also discuss various ensembling strategies. Our framework is substantially sample efficient, computationally inexpensive and is seen to outperform state of the art (SOTA) approaches
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.