A5044488056- Transportation and Mobility Innovations
- Simulation Techniques and Applications
- Traffic control and management
- Real-time simulation and control systems
- Advanced Vision and Imaging
- Autonomous Vehicle Technology and Safety
- Robotics and Sensor-Based Localization
- Robotic Path Planning Algorithms
- Computer Graphics and Visualization Techniques
- Modeling and Simulation Systems
- Dynamics and Control of Mechanical Systems
- Advanced Optical Sensing Technologies
- Image Processing Techniques and Applications
- Vehicle Dynamics and Control Systems
- Reinforcement Learning in Robotics
- Multi-Agent Systems and Negotiation
- Robot Manipulation and Learning
- Vehicle emissions and performance
- Hydraulic and Pneumatic Systems
- Assembly Line Balancing Optimization
- Advanced Image and Video Retrieval Techniques
- Advanced Image Processing Techniques
- Aerospace and Aviation Technology
- Electric and Hybrid Vehicle Technologies
- Metallurgy and Material Science
University of Wisconsin–Madison
2017-2024
Jet Propulsion Laboratory
2024
The last five years marked a surge in interest for and use of smart robots, which operate dynamic unstructured environments might interact with humans. We posit that well-validated computer simulation can provide virtual proving ground many cases is instrumental understanding safely, faster, at lower costs, more thoroughly how the robots future should be designed controlled safe operation improved performance. Against this backdrop, we discuss help robotics, barriers currently prevent its...
Simulation can be an important tool in evaluating autonomous vehicle performance. In a virtual proving ground, challenging and highly unlikely scenarios created novel control algorithms tested with no safety concerns relatively little cost. A critical component of any ground is sensor simulation; i.e., the ability to produce realistic information inside simulated environment. This paper overviews several models methods for generating, through simulation, GPS, IMU, lidar data. While current...
Abstract Computer simulation can be a useful tool when designing robots expected to operate independently in unstructured environments. In this context, one needs simulate the dynamics of robot’s mechanical system, environment which robot operates, and sensors facilitate perception environment. Herein, we focus on sensing task by presenting virtual framework built alongside an open-source, multi-physics platform called Chrono. This supports camera, lidar, GPS, IMU simulation. We discuss...
View Video Presentation: https://doi.org/10.2514/6.2022-1433.vid Perception plays a key role in autonomous and semi-autonomous planetary exploration vehicles. For instance, landers can use computer vision techniques for identifying safe landing locations, aerial vehicles cameras as navigation sensors, rovers them localization hazard detection. Engineering simulations of such systems require the accurate modeling perception sensors simulating autonomy scenarios. In addition, landers, ground...
Simulation is increasingly important in the development and testing of robots autonomous vehicles as it opens door for candidate navigation, perception, sensor fusion algorithms to be expeditiously probed complex safety-critical scenarios. As most make heavy use cameras perceive their surroundings, camera modeling becomes a prerequisite successful simulation these agents. This contribution outlines context which models are used; provides component-by-component analysis image acquisition...
We describe a software framework and hardware platform used in tandem for the design analysis of robot autonomy algorithms simulation reality. The software, which is open source, containerized, operating system (OS) independent, has three main components: ROS 2 interface to C++ vehicle (Chrono), provides high-fidelity wheeled/tracked sensor simulation; basic 2-based stack algorithm testing; and, development ecosystem enables visualization, hardware-in-the-loop experimentation perception,...
<title>ABSTRACT</title> <p>We describe a simulation environment that enables the design and testing of control policies for off-road mobility autonomous agents. The is demonstrated in conjunction with assessment reinforcement learning policy uses sensor fusion inter-agent communication to enable movement mixed convoys conventional vehicles. Policies learned on rigid terrain are shown transfer hard (silt-like) soft (snow-like) deformable terrains. enabling environment, which...
Abstract We describe a simulation environment that enables the design and testing of control policies for off-road mobility autonomous agents. The is demonstrated in conjunction with training assessment reinforcement learning policy uses sensor fusion interagent communication to enable movement mixed convoys human-driven vehicles. Policies learned on rigid terrain are shown transfer hard (silt-like) soft (snow-like) deformable terrains. described performs following: multivehicle multibody...
EELS-DARTS is a simulator designed for autonomy development and analysis of large degree freedom snake-like robots space exploration. A detailed description the design presented. This includes versatile underlying multibody dynamics representation used to model variety distinct snake robot configurations as well an anisotropic friction describing screw–ice interaction. Additional simulation components such graphics, importable terrain, joint controllers, perception are discussed. Methods...
This contribution is concerned with the topic of using simulation to understand behavior groups mutually interacting autonomous vehicles (AVs) or robots engaged in traffic/maneuvers that involve coordinated operation. We outline structure a multi-agent simulator called SYN-CHRONO and provide results pertaining its scalability ability run real-time scenarios humans loop. scalable multi-agent, high-fidelity environment whose purpose testing AV robot control strategies. Four main components...
Abstract We describe a simulation environment that enables the development and testing of control policies for off-road mobility autonomous agents. The is demonstrated in conjunction with design assessment reinforcement learning policy uses sensor fusion inter-agent communication to enable movement mixed convoys human-driven vehicles. Policies are learned on rigid terrain subsequently shown transfer successfully hard (silt-like) soft (snow-like) deformable terrains. enabling developed from...
We discuss an early version of open-source autonomous vehicle simulation framework whereby piloting control programs (PCPs) and response can be evaluated improved in a safe environment. Through the interaction hundreds avatar agents simulated environment, edge cases for self-driving vehicles analyzed thus accelerating research, development technology deployment. The Connected Autonomous Vehicle Emulator (CAVE) builds upon four foundational components: (i) physics engine dynamic support using...
We present an open-source framework called SYNCHRONO that enables one to use physics-based simulation gauge how collaborating robots work together in a variety of environments. Building on top the CHRONO dynamics engine [1], provides early support for operating off-road conditions, underwater, city environments, etc. In this contribution we focus autonomous vehicles (AVs), which represent but many identities "robot" can assume. context, draws template-based vehicle library enable shared-road...
The focus of this contribution is on camera simulation as it comes into play in simulating autonomous robots for their virtual prototyping. We propose a model validation methodology based the performance perception algorithm and context which measured. This approach different than traditional synthetic images, often done at pixel or feature level, tends to require matching pairs real images. Due high cost constraints acquiring paired proposed datasets that are not necessarily paired. Within...
Given the versatility of generative adversarial networks (GANs), we seek to understand benefits gained from using an existing GAN enhance simulated images and reduce sim-to-real gap. We conduct analysis in context simulating robot performance image-based perception. Specifically, quantify GAN's ability difference image perception robotics. Using semantic segmentation, analyze training testing, nominal enhanced simulation a city environment. As secondary application, consider use enhancing...
Modeling cameras for the simulation of autonomous robotics is critical generating synthetic images with appropriate realism to effectively evaluate a perception algorithm in simulation. In many cases though, simulated are produced by traditional rendering techniques that exclude or superficially handle processing steps and aspects encountered actual camera pipeline. The purpose this contribution quantify effect modifying model has on evaluated We investigate what happens if one ignores tied...