Optimal Control Learning in Noisy Environments
نویسنده
چکیده
This paper presents a motor control learning strategy in the presence of noise. Noise is important to consider as any real-world environment contain elements that cannot be accurately modeled. Learning is accomplished through analysis of control-response pairs. Policy parameters are sampled from multivariate normal distributions that slowly converge around the optimal policy, which can then be extracted. The algorithm is demonstrated in cannonball-launching and dart-throwing simulations implemented in the Box2D framework. Results show that optimal strategies can be obtained in the presence of noise. Moreover, the adding of noise allows for the detection of non-robust policies.
منابع مشابه
Learning from Demonstration in Static Environment and Generalizing to Dynamic Environments
Robot learning from demonstration has been successfully used, in industrial environments, to increase the speed and reduce the complexity of the programming phase. A major challenge, however, consists in enabling a robot to generalize task demonstrations to a complex dynamic environment. To tackle this problem, an approach has been proposed in [1], combining GMM/GMR and DMPs with optimal contro...
متن کاملOn the effect of low-quality node observation on learning over incremental adaptive networks
In this paper, we study the impact of low-quality node on the performance of incremental least mean square (ILMS) adaptive networks. Adaptive networks involve many nodes with adaptation and learning capabilities. Low-quality mode in the performance of a node in a practical sensor network is modeled by the observation of pure noise (its observation noise) that leads to an unreliable measurement....
متن کاملThe effects of traffic noise on memory and auditory-verbal learning in Persian language children
Background: Acoustic noise is one of the universal pollutants of modern society. Although the high level of noise adverse effects on human hearing has been known for many years, non-auditory effects of noise such as effects on cognition, learning, memory and reading, especially on children, have been less considered. Factors which have negative impact on these features can also have a negat...
متن کاملAddressing Environment Non-Stationarity by Repeating Q-learning Updates
Q-learning (QL) is a popular reinforcement learning algorithm that is guaranteed to converge to optimal policies in Markov decision processes. However, QL exhibits an artifact: in expectation, the effective rate of updating the value of an action depends on the probability of choosing that action. In other words, there is a tight coupling between the learning dynamics and underlying execution p...
متن کاملEffective Environmental Factors on Designing Productive Learning Environments
Educational spaces play an important role in enhancing learning productivity levels of society people as the most important places to human train. Considering the cost, time and energy spending on these spaces, trying to design efficient and optimized environment is a necessity. Achieving efficient environments requires changing environmental criteria so that they can have a positive impact on ...
متن کامل