A5070031738- Target Tracking and Data Fusion in Sensor Networks
- Robotics and Sensor-Based Localization
- Distributed Control Multi-Agent Systems
- Robotic Path Planning Algorithms
- Distributed Sensor Networks and Detection Algorithms
- Anomaly Detection Techniques and Applications
- Guidance and Control Systems
- Data Management and Algorithms
- Fault Detection and Control Systems
- Energy Efficient Wireless Sensor Networks
- Autonomous Vehicle Technology and Safety
- Indoor and Outdoor Localization Technologies
- Optimization and Search Problems
- Remote Sensing and LiDAR Applications
- Advanced Image and Video Retrieval Techniques
- Neural Networks and Applications
- Traffic Prediction and Management Techniques
- Software Testing and Debugging Techniques
- Human Pose and Action Recognition
- Time Series Analysis and Forecasting
- Human Mobility and Location-Based Analysis
- Gaussian Processes and Bayesian Inference
- 3D Surveying and Cultural Heritage
- Security in Wireless Sensor Networks
- Reinforcement Learning in Robotics
Temple University
2016-2025
Temple College
2022-2024
University of Science and Technology
2022
University of Pennsylvania
2012-2016
California University of Pennsylvania
2013
The use of aerial swarms to solve real-world problems has been increasing steadily, accompanied by falling prices and improving performance communication, sensing, processing hardware. commoditization hardware reduced unit costs, thereby lowering the barriers entry field swarm robotics. A key enabling technology for is family algorithms that allow individual members communicate allocate tasks amongst themselves, plan their trajectories, coordinate flight in such a way overall objectives are...
This article proposes a novel learning-based control policy with strong generalizability to new environments that enables mobile robot navigate autonomously through spaces filled both static obstacles and dense crowds of pedestrians. The uses unique combination input data generate the desired steering angle forward velocity: short history lidar data, kinematic about nearby pedestrians, subgoal point. is trained in reinforcement learning setting using reward function contains term based on...
In this paper we examine the problem of autonomously exploring and mapping an environment using a mobile robot. The robot uses graph-based SLAM system to perform represents map as occupancy grid. setting, must trade-off between new area complete task exploiting existing information maintain good localization. Selecting actions that decrease uncertainty while not significantly increasing robot's localization is challenging. We present novel information-theoretic utility function both...
This paper considers situations in which a team of mobile sensor platforms autonomously explores an environment to detect and localize unknown number targets. Individual sensors may be unreliable, failing objects within the field-of-view, returning false positive measurements clutter objects, being unable disambiguate true In this setting, data association is difficult. We utilize PHD filter for multitarget localization, simultaneously estimating their locations without need explicitly...
This paper proposes a distributed estimation and control algorithm that enables team of mobile robots to search for track an unknown number targets. These targets may be stationary or moving, the vary over time as enter leave area interest. The are equipped with sensors have finite field view experience false negative positive detections. use novel, formulation Probability Hypothesis Density (PHD) filter, which accounts limitations sensors, estimate positions then Lloyd's algorithm, has been...
This paper proposes an algorithm for driving a group of resource-constrained robots with noisy sensors to localize unknown number targets in environment, while avoiding hazards at positions that cause the fail. The is based upon analytic gradient mutual information target locations and measurements offers two primary improvements over previous algorithms [6], [13]. Firstly, it decentralized. follows from approximation fact robots' environmental have finite area influence. Secondly, allows be...
Magnetic anomaly detection (MAD) is an important problem in applications ranging from geological surveillance to military reconnaissance. MAD sensors detect local disturbances the magnetic field, which can be used existence of and estimate position buried, hidden, or submerged objects, such as ore deposits mines. These may experience false positive negative detections and, without prior knowledge targets, only determine proximity a target. The uncertainty sensors, coupled with lack even...
Accurately tracking dynamic targets relies on robots accounting for uncertainties in their own states to share information and maintain safety. The problem becomes even more challenging when there is an unknown time-varying number of the environment. In this paper we address by introducing four new distributed algorithms that allow large teams to: i) run prediction ii) update steps a recursive Bayesian multi- target tracker, iii) determine set local neighbors must exchange data, iv) data...
This paper introduces the normalized unused sensing capacity to measure amount of information that a sensor is currently gathering relative its theoretical maximum. quantity can be computed using entirely local and works for arbitrary models, unlike previous literature on subject. then used develop distributed coverage control strategy team heterogeneous sensors automatically balances load based current each member. algorithm validated in multi-target tracking scenario, yielding superior...
In this paper, we propose a distributed coverage control algorithm for mobile sensing networks that can account bounded uncertainty in the location of each sensor. Our is capable safely driving sensors towards areas high information distribution while having them maintain whole area interest. To do this, two novel variants Voronoi diagram. The first, convex uncertain (CUV) diagram, guarantees full search area. second, collision avoidance regions (CARs), guarantee collision-free motions...
This technical report is an extended version of the paper 'Cooperative Multi-Target Localization With Noisy Sensors' accepted to 2013 IEEE International Conference on Robotics and Automation (ICRA). addresses task searching for unknown number static targets within a known obstacle map using team mobile robots equipped with noisy, limited field-of-view sensors. Such sensors may fail detect subset visible or return false positive detections. These measurement sets are used localize Probability...
The Benchmark Autonomous Robot Navigation (BARN) Challenge took place at the 2022 IEEE International Conference on Robotics and Automation (ICRA), in Philadelphia, PA, USA. aim of challenge was to evaluate state-of-the-art autonomous ground navigation systems for moving robots through highly constrained environments a safe efficient manner. Specifically, task navigate standardized differential drive robot from predefined start location goal as quickly possible without colliding with any...
Target tracking is a fundamental problem in robotics research and has been the subject of detailed studies over years. In this paper, we generate data-driven target model from real-world dataset taxi motions. This includes motion, appearance, disappearance search area. Using model, introduce new formulation mobile based on mathematical concept random finite sets. allows for an unknown dynamic number targets with team robots. We show how to employ probability hypothesis density filter...
Abstract Distributed search and track is a canonical task for multi-robot systems, encompassing applications from environmental monitoring to disaster response surveillance. In many situations, the distribution of objects in area irregular, with some areas having high object densities while other have low densities. this paper, we formulate as multi-armed bandit problem propose novel distributed formulation Bernoulli Thompson sampling that enables robots share coarse global information...
This paper presents an automated robotic system for generating semantic maps of inventory in retail environments. In settings, are labeled stores where each discrete section shelving is assigned a department label describing the types products on that shelf. Starting from metric map store, robot autonomously extracts shelf boundaries, generates distance-optimal tour store to view every shelf, and follows while avoiding unmapped clutter moving people. The creates point cloud using data...