Model-free Deep Reinforcement Learning Algorithm Research Articles

Interaction-Aware Decision-Making for Automated Vehicles Using Social Value Orientation

Motion control algorithms in the presence of pedestrians are critical for the development of safe and reliable Autonomous Vehicles (AVs). Traditional motion control algorithms rely on manually designed decision-making policies which neglect the mutual interactions between AVs and pedestrians. On the other hand, recent advances in Deep Reinforcement Learning allow for the automatic learning of policies without manual designs. To tackle the problem of decision-making in the presence of pedestrians, the authors introduce a framework based on Social Value Orientation and Deep Reinforcement Learning (DRL) that is capable of generating decision-making policies with different driving styles. The policy is trained using state-of-the-art DRL algorithms in a simulated environment. A novel computationally-efficient pedestrian model that is suitable for DRL training is also introduced. We perform experiments to validate our framework and we conduct a comparative analysis of the policies obtained with two different model-free Deep Reinforcement Learning Algorithms. Simulations results show how the developed model exhibits natural driving behaviours, such as short-stopping, to facilitate the pedestrian's crossing.

Active flutter suppression for a flexible wing model with trailing-edge circulation control via reinforcement learning

Previous attempts at active flutter suppression have been based on driving the deflection of multiple pairs of discontinuous mechanical control surfaces. Here, we explore the effects of trailing-edge Circulation Control (CC) for flutter control on flexible wings. To avoid the problem that the nonlinear aeroelastic model is difficult to establish accurately, we trained a closed-loop control strategy based on the model-free deep reinforcement learning algorithm through aeroelastic wind tunnel testing. The results show that the strategy can intelligently select the appropriate jet intensity according to the real-time state of the flexible wing. The oscillation amplitude of flutter can be reduced by 92%. The air consumption required for unsteady CC to suppress flutter is reduced by 37% compared to steady CC. This study aims to provide an innovative control method and strategy for active flutter suppression of large aspect ratio flexible wings.

Deep Reinforcement Learning for Adaptive Learning Systems

The adaptive learning problem concerns how to create an individualized learning plan (also referred to as a learning policy) that chooses the most appropriate learning materials based on a learner’s latent traits. In this article, we study an important yet less-addressed adaptive learning problem—one that assumes continuous latent traits. Specifically, we formulate the adaptive learning problem as a Markov decision process. We assume latent traits to be continuous with an unknown transition model and apply a model-free deep reinforcement learning algorithm—the deep Q-learning algorithm—that can effectively find the optimal learning policy from data on learners’ learning process without knowing the actual transition model of the learners’ continuous latent traits. To efficiently utilize available data, we also develop a transition model estimator that emulates the learner’s learning process using neural networks. The transition model estimator can be used in the deep Q-learning algorithm so that it can more efficiently discover the optimal learning policy for a learner. Numerical simulation studies verify that the proposed algorithm is very efficient in finding a good learning policy. Especially with the aid of a transition model estimator, it can find the optimal learning policy after training using a small number of learners.

Retro-RL: Reinforcing Nominal Controller With Deep Reinforcement Learning for Tilting-Rotor Drones

Studies that broaden drone applications into complex tasks require a stable control framework. Recently, deep reinforcement learning (RL) algorithms have been exploited in many studies for robot control to accomplish complex tasks. Unfortunately, deep RL algorithms might not be suitable for being deployed directly into a real-world robot platform due to the difficulty in interpreting the learned policy and lack of stability guarantee, especially for a complex task such as a wall-climbing drone. This letter proposes a novel hybrid architecture that reinforces a nominal controller with a robust policy learned using a model-free deep RL algorithm. The proposed architecture employs an uncertainty-aware control mixer to preserve guaranteed stability of a nominal controller while using the extended robust performance of the learned policy. The policy is trained in a simulated environment with thousands of domain randomizations to achieve robust performance over diverse uncertainties. The performance of the proposed method was verified through real-world experiments and then compared with a conventional controller and the state-of-the-art learning-based controller trained with a vanilla deep RL algorithm.

Rapid Autonomous Vehicle Drifting with Deep Reinforcement Learning

Abstract: Now days all types of vehicles run on road are upgrading to automated self-driving techniques. Drifting is a complicated procedure for autonomous vehicle control which is used for high-speed sideslip cornering. It is difficult to drift a vehicle at a high speed of 80-120 km/hr by using traditional techniques like two-state single-track model and three-state singletrack models, which depend on the knowledge of tire or road forces that are not to be known accurately due to the real-world environmental complexity. Because of these inaccuracies in these parameters there will be poor control performance and may lead to road accidents. In this scenario, this project presents a strong drift control algorithm, which is based on the most recent model-free deep reinforcement learning algorithm called soft actor-critic (SAC). SAC is used to control front-wheel drive (FWD) vehicles to drive at high speed (80– 120 km/h), and to drift through sharp corners quickly and stability which in turn helps to prevent road accidents at the time of drifting. The efficiency of this algorithm is evaluated by conducting experiments on Trajectories dataset taken from Git-Hub repository, which consists of some road maps with reference drift trajectories. The SAC algorithm can also deal with vehicle types with various actual properties, like mass, tyre friction, and so on to show its notable generalization ability.

Autonomous robotic additive manufacturing through distributed model‐free deep reinforcement learning in computational design environments

The objective of autonomous robotic additive manufacturing for construction in the architectural scale is currently being investigated in parts both within the research communities of computational design and robotic fabrication (CDRF) and deep reinforcement learning (DRL) in robotics. The presented study summarizes the relevant state of the art in both research areas and lays out how their respective accomplishments can be combined to achieve higher degrees of autonomy in robotic construction within the Architecture, Engineering and Construction (AEC) industry. A distributed control and communication infrastructure for agent training and task execution is presented, that leverages the potentials of combining tools, standards and algorithms of both fields. It is geared towards industrial CDRF applications. Using this framework, a robotic agent is trained to autonomously plan and build structures using two model-free DRL algorithms (TD3, SAC) in two case studies: robotic block stacking and sensor-adaptive 3D printing. The first case study serves to demonstrate the general applicability of computational design environments for DRL training and the comparative learning success of the utilized algorithms. Case study two highlights the benefit of our setup in terms of tool path planning, geometric state reconstruction, the incorporation of fabrication constraints and action evaluation as part of the training and execution process through parametric modeling routines. The study benefits from highly efficient geometry compression based on convolutional autoencoders (CAE) and signed distance fields (SDF), real-time physics simulation in CAD, industry-grade hardware control and distinct action complementation through geometric scripting. Most of the developed code is provided open source.

Robotic architectural assembly with tactile skills: Simulation and optimization

Construction is an industry that could benefit significantly from automation, yet still relies heavily on manual human labor. Thus, we investigate how a robotic arm can be used to assemble a structure from predefined discrete building blocks autonomously. Since assembling structures is a challenging task that involves complex contact dynamics, we propose to use a combination of reinforcement learning and planning for this task. In this work, we take a first step towards autonomous construction by training a controller to place a single building block in simulation. Our evaluations show that trial-and-error algorithms that have minimal prior knowledge about the problem to be solved, so called model-free deep reinforcement learning algorithms, can be successfully employed. We conclude that the achieved results, albeit imperfect, serve as a proof of concept and indicate the directions for further research to enable more complex assemblies involving multiple building elements.

Data-driven battery operation for energy arbitrage using rainbow deep reinforcement learning

As the world seeks to become more sustainable, intelligent solutions are needed to increase the penetration of renewable energy. In this paper, the model-free deep reinforcement learning algorithm Rainbow Deep Q-Networks is used to control a battery in a microgrid to perform energy arbitrage and more efficiently utilise solar and wind energy sources. The grid operates with its own demand and renewable generation, as well as dynamic energy pricing from a real wholesale energy market. Four scenarios are tested including using demand and price forecasting produced with local weather data. The algorithm and its subcomponents are evaluated against an actor-critic method and a linear programming model with Rainbow able to outperform all other methods. This research shows the importance of using the distributional approach for reinforcement learning for complex environments, as well as how it can visualise and contextualise the agents’ behaviour for real-world applications.

Deep reinforcement learning methods for structure-guided processing path optimization

A major goal of materials design is to find material structures with desired properties and in a second step to find a processing path to reach one of these structures. In this paper, we propose and investigate a deep reinforcement learning approach for the optimization of processing paths. The goal is to find optimal processing paths in the material structure space that lead to target-structures, which have been identified beforehand to result in desired material properties. There exists a target set containing one or multiple different structures, bearing the desired properties. Our proposed methods can find an optimal path from a start structure to a single target structure, or optimize the processing paths to one of the equivalent target-structures in the set. In the latter case, the algorithm learns during processing to simultaneously identify the best reachable target structure and the optimal path to it. The proposed methods belong to the family of model-free deep reinforcement learning algorithms. They are guided by structure representations as features of the process state and by a reward signal, which is formulated based on a distance function in the structure space. Model-free reinforcement learning algorithms learn through trial and error while interacting with the process. Thereby, they are not restricted to information from a priori sampled processing data and are able to adapt to the specific process. The optimization itself is model-free and does not require any prior knowledge about the process itself. We instantiate and evaluate the proposed methods by optimizing paths of a generic metal forming process. We show the ability of both methods to find processing paths leading close to target structures and the ability of the extended method to identify target-structures that can be reached effectively and efficiently and to focus on these targets for sample efficient processing path optimization.

Aerial filming with synchronized drones using reinforcement learning

Usage of multiple drones is necessary for aerial filming applications to ensure redundancy. However, this could inevitably contribute to higher risks of collisions, especially when the number of drones increases. Hence, this motivates us to explore various autonomous flight formation control methods that have the potential to enable multiple drones to effectively track a specific target at the same time. In this paper, we designed a model-free deep reinforcement learning algorithm, which is mainly based on the Deep Recurrent Q-Network concept, for the aforementioned purposes. The proposed algorithm was expanded into single and multi-agent types that enable multiple drones tracking while maintaining formation and preventing collision. The involved rewards in these approaches are two-dimensional in nature and are dependent on the communication system. Using Microsoft AirSim simulator, a virtual environment that includes four virtual drones was developed for experimental purposes. A comparison was made among various methods during the simulations, and the results concluded that the recurrent, single-agent model is the most effective method, being 33% more effective than its recurrent, multi-agent counterparts. The poor performance of the non-recurrent, single-agent baseline model also suggests that the recurrent elements in the network are essential to enable desirable multiple-drones flight.

FlexPool: A Distributed Model-Free Deep Reinforcement Learning Algorithm for Joint Passengers and Goods Transportation

The growth in online goods delivery is causing a dramatic surge in urban vehicle traffic from last-mile deliveries. On the other hand, ride-sharing has been on the rise with the success of ride-sharing platforms and increased research on using autonomous vehicle technologies for routing and matching. The future of urban mobility for passengers and goods relies on leveraging new methods that minimize operational costs and environmental footprints of transportation systems. This article considers combining passenger transportation with goods delivery to improve vehicle-based transportation. We propose FlexPool: a distributed model-free deep reinforcement learning algorithm that jointly serves passengers & goods workloads by learning optimal dispatch policies from its interaction with the environment. The proposed algorithm pools passengers for a ride-sharing service and delivers goods using a multi-hop transit method. These flexibilities decrease the fleet's operational cost and environmental footprint while maintaining service levels for passengers and goods. The dispatching algorithm based on deep reinforcement learning is integrated with an efficient matching algorithm for passengers and goods. Through simulations on a realistic multi-agent urban mobility platform, we demonstrate that FlexPool outperforms other model-free settings in serving the demands from passengers & goods. FlexPool achieves 30% higher fleet utilization and 35% higher fuel efficiency in comparison to (i) model-free approaches where vehicles transport a combination of passengers & goods without the use of multi-hop transit, and (ii) model-free approaches where vehicles exclusively transport either passengers or goods.

Research on sports action training method based on generative confrontation network model and artificial intelligence

In order to improve the effect of sports movement training, this paper builds a sports movement training model based on artificial intelligence technology based on the generation of confrontation network model. Moreover, in order to achieve the combination of model and model-free deep reinforcement learning algorithm, this paper implements the model’s guidance and constraints on deep reinforcement learning algorithm from the perspective of reward value and behavior strategy and divides the model into two situations. In one case, the existing or manually established expert rules are used as model constraints, which is equivalent to online learning by experts. In another case, expert samples are used as model constraints, and an imitation learning method based on generative adversarial networks is introduced. Moreover, using expert samples as training data, the mechanism that the model is guided by the reward value is combined with the model-free algorithm by generating a confrontation network structure. Finally, this paper studies the performance of the model through experimental research. The research results show that the model constructed in this paper has a certain effect.

Energy-Efficient Mode Selection and Resource Allocation for D2D-Enabled Heterogeneous Networks: A Deep Reinforcement Learning Approach

Improving energy efficiency has shown increasing importance in designing future cellular system. In this work, we consider the issue of energy efficiency in D2D-enabled heterogeneous cellular networks. Specifically, communication mode selection and resource allocation are jointly considered with the aim to maximize the energy efficiency in the long term. And an Markov decision process (MDP) problem is formulated, where each user can switch between traditional cellular mode and D2D mode dynamically. We employ deep deterministic policy gradient (DDPG), a model-free deep reinforcement learning algorithm, to solve the MDP problem in continuous state and action space. The architecture of proposed method consists of one actor network and one critic network. The actor network uses deterministic policy gradient scheme to generate deterministic actions for agent directly, and the critic network employs value function based Q networks to evaluate the performance of the actor network. Simulation results show the convergence property of proposed algorithm and the effectiveness in improving the energy efficiency in a D2D-enabled heterogeneous network.

Towards safe reinforcement-learning in industrial grid-warehousing

Reinforcement learning has shown to be profoundly successful at learning optimal policies for simulated environments using distributed training with extensive compute capacity. Model-free reinforcement learning uses the notion of trial and error, where the error is a vital part of learning the agent to behave optimally. In mission-critical, real-world environments, there is little tolerance for failure and can cause damaging effects on humans and equipment. In these environments, current state-of-the-art reinforcement learning approaches are not sufficient to learn optimal control policies safely.On the other hand, model-based reinforcement learning tries to encode environment transition dynamics into a predictive model. The transition dynamics describes the mapping from one state to another, conditioned on an action. If this model is accurate enough, the predictive model is sufficient to train agents for optimal behavior in real environments.This paper presents the Dreaming Variational Autoencoder (DVAE) for safely learning good policies with a significantly lower risk of catastrophes occurring during training. The algorithm combines variational autoencoders, risk-directed exploration, and curiosity to train deep-q networks inside ”dream” states. We introduce a novel environment, ASRS-Lab, for research in the safe learning of autonomous vehicles in grid-based warehousing. The work shows that the proposed algorithm has better sample efficiency with similar performance to novel model-free deep reinforcement learning algorithms while maintaining safety during training.

High-Speed Autonomous Drifting With Deep Reinforcement Learning

Drifting is a complicated task for autonomous vehicle control. Most traditional methods in this area are based on motion equations derived by the understanding of vehicle dynamics, which is difficult to be modeled precisely. We propose a robust drift controller without explicit motion equations, which is based on the latest model-free deep reinforcement learning algorithm soft actor-critic. The drift control problem is formulated as a trajectory following task, where the errorbased state and reward are designed. After being trained on tracks with different levels of difficulty, our controller is capable of making the vehicle drift through various sharp corners quickly and stably in the unseen map. The proposed controller is further shown to have excellent generalization ability, which can directly handle unseen vehicle types with different physical properties, such as mass, tire friction, etc.

Optimal energy management strategies for energy Internet via deep reinforcement learning approach

This paper investigates the energy management problem in the field of energy Internet (EI) with interdisciplinary techniques. The concept of EI has been proposed for a while. However, there still exist many fundamental and technical issues that have not been fully investigated. In this paper, a new energy regulation issue is considered based on the operational principles of EI. Multiple targets are considered along with constraints. Then, the practical energy management problem is formulated as a constrained optimal control problem. Notably, no explicit mathematical model for power of renewable power generation devices and loads is utilized. Due to the complexity of this problem, conventional methods appear to be inapplicable. To obtain the desired control scheme, a model-free deep reinforcement learning algorithm is applied. A practical solution is obtained, and the feasibility as well as the performance of the proposed method are evaluated with numerical simulations.