A5040054567- Robotic Path Planning Algorithms
- Autonomous Vehicle Technology and Safety
- Robotics and Sensor-Based Localization
- Video Surveillance and Tracking Methods
- Academic and Historical Perspectives in Psychology
- Advanced Neural Network Applications
- Human-Automation Interaction and Safety
- Quality and Safety in Healthcare
- Human Pose and Action Recognition
- Reinforcement Learning in Robotics
- Evacuation and Crowd Dynamics
- AI-based Problem Solving and Planning
- Cultural Differences and Values
- Engineering Education and Curriculum Development
- Social Robot Interaction and HRI
- Technology and Human Factors in Education and Health
- Evaluation of Teaching Practices
- Air Traffic Management and Optimization
- Aerospace and Aviation Technology
- Neural Networks and Applications
- Behavioral and Psychological Studies
- Advanced Text Analysis Techniques
- Advanced Malware Detection Techniques
- Design Education and Practice
- UAV Applications and Optimization
ETH Zurich
2015-2020
Dynamic Systems (United States)
2015
SAIL LABS Technology (Austria)
2008-2013
La Salle University
1965-1976
We study the use of kinematic and dynamic vehicle models for model-based control design used in autonomous driving. In particular, we analyze statistics forecast error these two by using experimental data. addition, effect discretization on error. results first part to motivate a controller an model predictive (MPC) simple bicycle model. The proposed approach is less computationally expensive than existing methods which tire models. Moreover it can be implemented at low speeds where become...
Learning from demonstration for motion planning is an ongoing research topic. In this paper we present a model that able to learn the complex mapping raw 2D-laser range findings and target position required steering commands robot. To our best knowledge, work presents first approach learns target-oriented end-to-end navigation robotic platform. The supervised training based on expert demonstrations generated in simulation with existing planner. We demonstrate learned directly transferable...
This letter presents a case study of learning-based approach for target-driven mapless navigation. The underlying navigation model is an end-to-end neural network, which trained using combination expert demonstrations, imitation learning (IL) and reinforcement (RL). While RL IL suffer from large sample complexity the distribution mismatch problem, respectively, we show that leveraging prior demonstrations pretraining can reduce training time to reach at least same level performance compared...
Abstract This paper presents the algorithms and system architecture of an autonomous racecar. The introduced vehicle is powered by a software stack designed for robustness, reliability, extensibility. To autonomously race around previously unknown track, proposed solution combines state art techniques from different fields robotics. Specifically, perception, estimation, control are incorporated into one high‐performance complex robotic system, developed AMZ Driverless ETH Zürich, finished...
This paper reports on a data-driven, interaction-aware motion prediction approach for pedestrians in environments cluttered with static obstacles. When navigating such workspaces shared humans, robots need accurate predictions of the surrounding pedestrians. Human navigation behavior is mostly influenced by their and obstacles vicinity. In this we introduce new model based Long-Short Term Memory (LSTM) neural networks, which able to learn human from demonstrated data. To best our knowledge,...
This paper reports on a data-driven motion planning approach for interaction-aware, socially-compliant robot navigation among human agents. Autonomous mobile robots navigating in workspaces shared with agents require techniques providing seamless integration and smooth such. Smooth mixed scenarios calls two abilities of the robot: predicting actions others acting predictably them. The former requirement requests trainable models agent behaviors order to accurately forecast their future,...
This paper reports on a data-driven motion planning approach for interaction-aware, socially-compliant robot navigation among human agents. Autonomous mobile robots navigating in workspaces shared with agents require techniques providing seamless integration and smooth such. Smooth mixed scenarios calls two abilities of the robot: predicting actions others acting predictably them. The former requirement requests trainable models agent behaviors order to accurately forecast their future,...
This paper presents a novel approach for local 3D environment representation autonomous unmanned ground vehicle (UGV) navigation called On Visible Point Clouds Mesh (OVPC Mesh). Our represents the surrounding of robot as watertight mesh generated from point cloud data in order to represent free space robot. It is conservative estimation and provides desirable trade-off between precision computational efficiency, without having discretize into fixed grid size. experiments analyze usability...
In autonomous racing, vehicles operate close to the limits of handling and a sensor failure can have critical consequences. To limit impact such failures, this paper presents redundant perception state estimation approaches developed for an race car. Redundancy in is achieved by estimating color position track delimiting objects using two modalities independently. Specifically, learning-based are used generate pose estimates, from LiDAR camera data respectively. The inputs fused particle...
In this paper we discuss the role of Open Source Intelligence (OSINT) in Disaster Management. particular present use Sail Labs Media Mining System context disaster relief operations and samples to point out advantages strengths MM-System. Future challenges research further development field are addressed towards end paper.
Identification of the specific contribution training system variables (e.g., simulator components, mixes, curriculum utilization strategies) will ultimately lead to more efficient for all tasks. An interest in acquisition piloting skills has led a number researchers question relative transfer simulated instrument and contact flight. In this study, we examined from Device 2F101, operational flight trainer T-2C aircraft, actual level speed change standard-rate turn maneuvers. Trainees were...
Autonomous driving and electric vehicles are nowadays very active research development areas. In this paper we present the conversion of a standard Kyburz eRod into an autonomous vehicle that can be operated in challenging environments such as Swiss mountain passes. The overall hardware software architectures described detail with special emphasis on sensor requirements for operating partially structured environments. Furthermore, design process itself finalized system architecture...
Visual localization and mapping is a crucial capability to address many challenges in mobile robotics. It constitutes robust, accurate cost-effective approach for local global pose estimation within prior maps. Yet, highly dynamic environments, like crowded city streets, problems arise as major parts of the image can be covered by objects. Consequently, visual odometry pipelines often diverge systems malfunction detected features are not consistent with precomputed 3D model. In this work, we...
30 operational tasks of psychology professors were scaled by the 14 students and 10 faculty same department using an adaptation pair comparisons method. The stimuli, in form statements classroom teacher activities, compared two at a time Ss who made judgments terms amount similarity between activities. These then transformed into correlation coefficients factor analysed. Factorial congruence student was demonstrated. activities urban college may be described 8 basic dimensions: “Knowledge...
We describe the Sail Labs Media Mining System which is capable of processing vast amounts data typically gathered from open sources in unstructured form. The are processed by a set components and output produced MPEG7 format. origin kind input may be as diverse satellite receivers monitoring TV stations or textual web-pages RSS-feeds. A sequence steps analyzing audio, video content carried out. resulting made available for search retrieval, analysis visualization on next generation Server....
This work presents a case study of learning-based approach for target driven map-less navigation. The underlying navigation model is an end-to-end neural network which trained using combination expert demonstrations, imitation learning (IL) and reinforcement (RL). While RL IL suffer from large sample complexity the distribution mismatch problem, respectively, we show that leveraging prior demonstrations pre-training can reduce training time to reach at least same level performance compared...
4 multidimensional scaling analyses based on similarity ratings among 30 operational statements of college-level psychology teaching activities (stimuli) using samples students and teachers from the USA Germany as raters resulted in 7 factors common varying degrees to both cultures one each culture with no counterpart other. The paired German-American ordered most least agreement accordance proportion their highly loaded stimuli held by corresponding across were: Teacher (Lecturing)...