Projects

• Home
• Research
  • Human Gesture Analysis
  • Smart Environments
  • Earlier Projects
• People
• Publications
• Courses
• Contact

Vision-based Localization

• Beacon-assisted Localization

  This work addresses the sensor network localization problem in a novel way, in that visual observations made by image sensors from a beacon agent are used to estimate network node coordinates. Based on visual observations of a moving beacon by the network nodes, the proposed technique employs simple image processing functions to produce equations that contain the node positions and orientation angles as the unknown parameters. Observations made at the nodes relate the position of the observed beacon to the physical coordinates of the node via the mapped position of the object in the node's image plane. Multiple observations by a network node from a moving beacon with known coordinates result in a system of equations with a rank-deficient matrix. Hence, the solution for the desired node coordinates lies in the null space of the data matrix. The proposed algorithms are based on in-node processing and hence are scalable to large networks. This approach is decentralized and no collabotion between the network nodes is ncessary. Hence, the network nodes can stay passive and only receive the beacon broadcasts during localization.
  Application areas in which the image sensor node location and orientation information can be used include target detection and tracking, robot tracking and control, estimation of traffic speed in roadways, and implementing geographic routing schemes for wireless sensor networks.

• Target-based Localization

  A collaborative vision-based technique is develop for localizing the nodes of a surveillance network based on observations of a non-cooperative moving target. The proposed method employs lightweight in-node image processing and limited data exchange between the nodes to determine the positions and orientations of the nodes participating in synchronized observations of the target. A node with an opportunistic observation of a passing target broadcasts a synchronizing packet and triggers image capture by its neighbors. In the cluster of participating nodes, the triggering node and a helper node define a relative coordinate system. Once a small number of joint observations of the target are made by the nodes, the model allows for a decentralized or a clusterbased solution for the localization problem. No images are transferred between the network nodes for the localization task, making the proposed method efficient and scalable.

• Robot-assisted Topology Discovery

  A localization solution focusing on network topology discovery for vision-enabled wireless sensor networks is developed in this project. A controllable robot is introduced to assist the localization process of image sensors deployed on the ceiling with image planes parallel to the ground. The scenarios in which the robot does not know its global coordinates is considered. Two cases where the sensors have overlapping and non-overlapping fields of view (FOVs) are investigated. In order to implement the discovery algorithms for these two different cases, a forest structure is introduced to represent the topology of the network. We consider the collection of sensors with overlapping FOVs as a tree in the forest. The robot searches for nodes in each tree through boundary patrolling, while it searches for other trees by a radial pattern motion.