Communicating UAVs (Unmanned Aerial Vehicles) are promising in terms of applications, but require to address numerous research issues. In this paper, we study the impact of the quality of Wi-Fi communications on the b...
详细信息
ISBN:
(纸本)9798350369458;9798350369441
Communicating UAVs (Unmanned Aerial Vehicles) are promising in terms of applications, but require to address numerous research issues. In this paper, we study the impact of the quality of Wi-Fi communications on the behavior of remotely controlled UAVs. To this end, we design an experimental platform composed of PX4 Vision UAVs, a Motion Capture system, and a tool to generate mission traffic from the ground station. By defining different scenarios and different traffic configurations, we conduct a set of experiments. This allows us to analyze how the presence of mission traffic impacts the stability of the reception of the control traffic, and how, in turn, it impacts the behavior of the UAVs.
Autonomous cooperative planning (ACP) is a promising technique to improve the efficiency and safety of multi-vehicle interactions for future intelligent transportation systems. However, realizing robust ACP is a chall...
详细信息
ISBN:
(纸本)9798350377712;9798350377705
Autonomous cooperative planning (ACP) is a promising technique to improve the efficiency and safety of multi-vehicle interactions for future intelligent transportation systems. However, realizing robust ACP is a challenge due to the aggregation of perception, motion, and communication uncertainties. This paper proposes a novel multi-uncertainty aware ACP (MUACP) framework that simultaneously accounts for multiple types of uncertainties via regularized cooperative model predictive control (RC-MPC). The regularizers and constraints for perception, motion, and communication are constructed according to the confidence levels, weather conditions, and outage probabilities, respectively. The effectiveness of the proposed method is evaluated in the Car Learning to Act (CARLA) simulation platform. Results demonstrate that the proposed MUACP efficiently performs cooperative formation in real time and outperforms other benchmark approaches in various scenarios under imperfect knowledge of the environment.
Constrained reinforcement learning is of great practical interest due to the pervasive existence of constraints in applications. Beyond the typical constraints directly on the state space, simulation-based constraints...
详细信息
ISBN:
(纸本)9798350355376;9798350355369
Constrained reinforcement learning is of great practical interest due to the pervasive existence of constraints in applications. Beyond the typical constraints directly on the state space, simulation-based constraints are harder to address, due to noisy and time consuming evaluation on both the performance and the feasibility of a policy. We consider this important problem in this work and make two contributions. First, we develop an algorithm based on Q-learning that iteratively improves the performance and the feasibility of a policy and show its global convergence. Second, for online learning we develop an algorithm to control the sampling among the action space, which is shown to asymptotically maximize the probability of correctly selecting the best feasible action.
The increasing need for effective vertical transit systems in present-day areas has led to the creation of clever elevator system solutions. In order to improve floor-level prioritizing, this study presents a Smart El...
详细信息
This present paper suggests an intelligent nonlinear controller based on the powerful law which is the sliding mode (SM) combined with the fractional calculus (FC) and optimized using the genetic algorithm (GA). This ...
详细信息
Robotics is an emerging field of interest among students to explore attractive tasks with their ideas. For engineering design, analytical observations are mandatory for the identification of further improvements in th...
详细信息
This paper presents a real-time robotic movement control system (RMCS) utilizing hand gesture recognition through a Cascaded CNN-SVM approach. The system was designed to enhance human-robot interaction by allowing use...
详细信息
The street light management system, which integrates new technologies primarily used for intelligent and adaptive-weather street lighting, is a new intelligent system that runs automatically. The traditional street li...
详细信息
The proceedings contain 176 papers. The topics discussed include: removing haze influence from remote sensing images captured with airborne visible/ infrared imaging spectrometer by cascaded fusion of DCP, GF, LCC wit...
ISBN:
(纸本)9781728185293
The proceedings contain 176 papers. The topics discussed include: removing haze influence from remote sensing images captured with airborne visible/ infrared imaging spectrometer by cascaded fusion of DCP, GF, LCC with AHE;a heuristic approach for text classification with ontology: a review;effectuating supervised machine learning techniques for multiclass classification of problematic internet and mobile usage;smart houses with the application of energy management system & smart grid;USB rubber ducky detection by using heuristic rules;defense against frequency analysis in elliptic curve cryptography using K-means clustering;efficient detection Of SQL injection attack (SQLIA) using pattern-based neural network model;effectuating supervised machine learning techniques for multiclass classification of problematic internet and mobile usage;and comparing ROC curve based thresholding methods in online transactions fraud detection system using deep learning.
The advent of tactile sensors in robotics has sparked many ideas on how robots can leverage direct contact measurements of their environment interactions to improve manipulation tasks. An important line of research in...
详细信息
ISBN:
(纸本)9798350377712;9798350377705
The advent of tactile sensors in robotics has sparked many ideas on how robots can leverage direct contact measurements of their environment interactions to improve manipulation tasks. An important line of research in this regard is grasp force control, which aims to manipulate objects safely by limiting the amount of force exerted on the object. While prior works have either hand-modeled their force controllers, employed model-based approaches, or not shown sim-to-real transfer, we propose a model-free deep reinforcement learning approach trained in simulation and then transferred to the robot without further fine-tuning. We, therefore, present a simulation environment that produces realistic normal forces, which we use to train continuous force control policies. A detailed evaluation shows that the learned policy performs similarly or better than a hand-crafted baseline. Ablation studies prove that the proposed inductive bias and domain randomization facilitate sim-to-real transfer. Code, models, and supplementary videos are available on https://***/view/rl-force-ctrl
暂无评论