Reinforcement Learning: Mdp Applied to Autonomous Navigation
نویسنده
چکیده
The problem of autonomous vehicle navigation between lanes, around obstacles and towards a short term goal can be solved using Reinforcement Learning. The multi-lane road ahead of a vehicle may be represented by a Markov Decision Process (MDP) grid-world containing positive and negative rewards, allowing for practical computation of an optimal path using either value iteration (VI) or policy iteration (PI).
منابع مشابه
A Generalized Reinforcement-Learning Model: Convergence and Applicationa
Reinforcement learning is the process by which an autonomous agent uses its experience interacting with an environment to improve its behavior. The Markov decision process (mdp) model is a popular way of formalizing the reinforcement-learning problem, but it is by no means the only way. In this paper, we show how many of the important theoretical results concerning reinforcement learning in mdp...
متن کاملCombining manual feedback with subsequent MDP reward signals for reinforcement learning
As learning agents move from research labs to the real world, it is increasingly important that human users, including those without programming skills, be able to teach agents desired behaviors. Recently, the tamer framework was introduced for designing agents that can be interactively shaped by human trainers who give only positive and negative feedback signals. Past work on tamer showed that...
متن کاملEfficient Human Following Using Reinforcement Learning
In this paper, we present an approach that relies on machine learning techniques to follow people efficiently during robotic assistance tasks, in which the robot is mainly interested in reaching the final navigation goal of the human. People can perform unexpected actions during navigation, which can lead to inefficient trajectories to the target destination (ex: answer land-line phones ... etc...
متن کاملNeuro-mimetic Navigation Systems: a Computational Model of the Rat Hippocampus Neuro-mimetic Navigation Systems: a Computational Model of the Rat Hippocampus
We propose a bio-inspired approach to autonomous navigation based on some of the components that rats use for navigation. A spatial model of the environment is constructed by unsupervised Hebbian learning. The representation consists of a population of localized overlapping place elds, modeling place cell activity in the rat Hippocampus. Place elds are established by extracting spatio-temporal ...
متن کاملAutonomous Navigation in Partially Observable Environments Using Hierarchical Q-Learning
A self-learning adaptive flight control design allows reliable and effective operation of flight vehicles in a complex environment. Reinforcement Learning provides a model-free, adaptive, and effective process for optimal control and navigation. This paper presents a new and systematic approach combining Q-learning and hierarchical reinforcement learning with additional connecting Q-value funct...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2018