Multirobot Cooperative Learning for Predator Avoidance

La, Hung Manh; Lim, Ronny Salim; Sheng, Weihua

doi:10.1109/tcst.2014.2312392

Cited by 128 publications

(70 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…To ensure a stable training, in our experiemnt we use a purely exploratory policy for the first 10000 time steps then begin to train the networks. (d) The training of the critic network involves the computation of Q-value error δ j and the gradient ∇ ω q(s j , u j |ω) according to (18) and (23). This is easy to implement using the back propagation algorithms.…”

Section: A Implementation Detailsmentioning

confidence: 99%

“…Under the dynamic programming framework, the RL aims to solve Markov decision process (MDP) problems by solely using training samples of the decision process. It is a model-free method, and has been applied successfully for the robotic control, including the path planning for a mobile robot [15] or multirobot [16], the vision-based robotic motion control [17], the multirobot cooperation [18], the high-precision motion control [19], the autonomous underwater vehicle (AUV) [20], etc.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Setpoint Tracking for the Suspension System of Medium-Speed Maglev Trains via Reinforcement Learning

Zhao

You

Song

et al. 2019

2019 IEEE 15th International Conference on Control and Automation (ICCA)

View full text Add to dashboard Cite

The suspension regulation is critical to the operation of medium-low-speed maglev trains (mlsMTs). Due to uncertain environment, strong disturbances and high nonlinearity of the system dynamics, this problem cannot be well solved by most of the model-based controllers. In this paper, we propose a model-free controller by reformulating it as a continuous-state, continuous-action Markov decision process (MDP) with unknown transition probabilities. With the deterministic policy gradient and neural network approximation, we design reinforcement learning (RL) algorithms to solve the MDP and obtain a statefeedback controller by using sampled data from the suspension system. To further improve its performance, we adopt a double Q-learning scheme for learning the regulation controller. We illustrate that the proposed controllers outperform the existing PID controller with a real dataset from the mlsMT in Changsha, China and is even comparable to model-based controllers, which assume that the complete information of the model is known, via simulations.

show abstract

Section: A Implementation Detailsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Setpoint Tracking for the Suspension System of Medium-Speed Maglev Trains via Reinforcement Learning

Zhao

You

Song

et al. 2019

2019 IEEE 15th International Conference on Control and Automation (ICCA)

View full text Add to dashboard Cite

show abstract

“…Multi-robot systems often require forms of cooperation, collaboration or co-ordination among the entities, such as collision-avoidance (Remmersmann et al, 2010) or flocking control (La et al, 2015), and for those inter-robot interaction is needed. Robots may possess various levels of cognition; see for example (Celentano, 2014).…”

Section: Control Of Cognitive Multi-robot Systemsmentioning

confidence: 99%

Multi-robot Systems, Machine-Machine and Human-Machine Interaction, and Their Modelling

Celentano

Röning

2016

Proceedings of the 8th International Conference on Agents and Artificial Intelligence

View full text Add to dashboard Cite

Abstract:The control of multi-agent systems, including multi-robot systems, requires some level of context and environment awareness as well as interaction among the interworked cognitive entities, whether they are artificial or natural. Proper specification of the cognitive functionalities and of the corresponding interfaces helps in achieving the capability to reach interoperability across different operational domains, and to reuse the system design across different application domains. The model for interworking cognitive entities presented in this article, which includes explicitly interworking capabilities, is applied to two major classes of interaction in multi-robot systems. Being the model inspired by both artificial and natural systems, makes it suitable for both machine-machine and human-machine interaction.

show abstract

“…Q-learning methods have been applied on a variety of tasks by autonomous robots [1], and much research has been done in this field starting many years ago [2], with some work specific to continuous action spaces [3]- [6] and others on discrete action spaces [7]. Reinforcement Learning (RL) has been applied to locomotion [8] [9] and also to manipulation [10], [11].…”

Section: Introductionmentioning

confidence: 99%

Deep Reinforcement Learning Using Genetic Algorithm for Parameter Optimization

Sehgal¹,

La²,

Louis³

et al. 2019

2019 Third IEEE International Conference on Robotic Computing (IRC)

View full text Add to dashboard Cite

Reinforcement learning (RL) enables agents to take decision based on a reward function. However, in the process of learning, the choice of values for learning algorithm parameters can significantly impact the overall learning process. In this paper, we use a genetic algorithm (GA) to find the values of parameters used in Deep Deterministic Policy Gradient (DDPG) combined with Hindsight Experience Replay (HER), to help speed up the learning agent. We used this method on fetchreach, slide, push, pick and place, and door opening in robotic manipulation tasks. Our experimental evaluation shows that our method leads to better performance, faster than the original algorithm.

show abstract

Multirobot Cooperative Learning for Predator Avoidance

Cited by 128 publications

References 32 publications

Setpoint Tracking for the Suspension System of Medium-Speed Maglev Trains via Reinforcement Learning

Setpoint Tracking for the Suspension System of Medium-Speed Maglev Trains via Reinforcement Learning

Multi-robot Systems, Machine-Machine and Human-Machine Interaction, and Their Modelling

Deep Reinforcement Learning Using Genetic Algorithm for Parameter Optimization

Contact Info

Product

Resources

About