RRC

SLAM Pose-graph Robustification

The Simultaneous Localization and Mappingproblem (SLAM) in robotics is typically modeled as a dyadic graph of relative pose measurements taken by the robot. The graph nodes store the values representing the absolute pose ofthe robot at a given point of time. An edge connecting two nodes represents robot movement and it stores the measurements taken by the robot sensor while moving between two nodes. Theobjective of the SLAM problem is to find the optimal global measurements best satisfying the noisy relative measurements Robust kernels which are less sensitive to outliers are used to deal with noise and outlier measurements. However, robust kernels tend to be dependent on initialization and can fail as the number of outliers increase. Therefore, it’s important to identify and prune the outlier (noisy) measurements represented by incorrect loop closure edges for an accurate pose estimate. We propose a multi-scale Heat-Kernel analysis based loop closure edge pruning algorithm for the SLAM graph. We show that compared to other pruning algorithms, our algorithm has a substantially higher precision and recall when compared and is able to handle a large amount of outlier measurements.

Related Publications:

Reactive Collision Avoidance for Multiple Robots

Reactive Collision avoidance for non-holonomic robots is a challenging task because of the restrictions in the space of achievable velocities. The complexity increases further when multiple non-holonomic robots are operating in tight/cluttered spaces. The present work presents a framework specially carved out for such situations. But at the same time can be easily appended with any existing collision avoidance framework. At the crux of the methodology is the concept of non-linear time scaling which allows robots to reactively accelerate/de-accelerate without altering the geometric path. The framework introduced is completely independent of the robot kinematics and dynamics. As such it can be applied to any ground or aerial robot. Through this concept the collision avoidance is framed as a problem of choosing appropriate scaling transformations. This work present a ”scaled” variant of the collision cone concept which automatically induces distributiveness among robots. The efficacy of the proposed work is demonstrated through simulations of both ground as well as UAVs.

Related Publications:

Heterogeneous UGV-MAV Exploration

This work presents a novel exploration strategy for coordinated exploration between unmanned ground vehicles(UGV) and micro-air vehicles (MAV). The exploration is modeled as an Integer Programming (IP) optimization problem and the allocation of the vehicles(agents) to frontier locations is modeled using binary variables. The formulation is also studied for distributed system, where agents are divided into multiple teams using graph partitioning. Optimization seamlessly integrates several practical constraints that arise in exploration between such heterogeneous agents and provides an elegant solution for assigning task to agents. We have also presented comparison with previous methods based on distance traversed and computational time to signify advantages of presented method. We also show practical realization of such an exploration where an UGV-MAV team efficiently builds a map of an indoor environment.

Related Publications:

UGV-UAV Co-ordination

This project aims at building a co-operative system consisting of a ground robot and a micro aerial vehicle(MAV). The Ground robot is used to create an accurate two-dimensional map of the surroundings as well as localize the ground robot. The MAV hovers over and follows the ground robot and finds new features to add to the map. The video feed from the MAV can be used to find new visual features and track old features and consequently map them into their world 3-dimensional coordinates. This visual based sparse point cloud, and the accurate and dense yet two-dimensional laser scan can then be fused to give rise to a better and more accurate 3-dimensional map of the surroundings.

Constrained Exploration

This project aims at building a co-operative system consisting of a ground robot and a micro aerial vehicle(MAV). The Ground robot is used to create an accurate two-dimensional map of the surroundings as well as localize the ground robot. The MAV hovers over and follows the ground robot and finds new features to add to the map. The video feed from the MAV can be used to find new visual features and track old features and consequently map them into their world 3-dimensional coordinates. This visual based sparse point cloud, and the accurate and dense yet two-dimensional laser scan can then be fused to give rise to a better and more accurate 3-dimensional map of the surroundings.

Related Publications:

Videos:

Terrain Exploration and Coverage with Multiple Robots

We are working on multi-robot coverage and exploration of both 2D & 3D terrains. We have come up with coupled metrics, strategies for improved performance, and in case of 3D, a framework for efficient terrain exploration. In general, our work deals with metrics, visibility models, communication constraints and motion strategies for reduced time paths and reduced distance paths.

Related Publications:

Intelligent Traffic Management System

We are working on methodologies for coordination of multiple robotic agents moving from one location to another in an environment embedded with sensor motes or the Intersection Agents (IAs). Sensor motes placed at strategic locations such as intersections coordinate robots in a way as to minimize the congestion, thus ensuring the continuous flow of robot traffic.

Related Publications:

Multi-Robot active localization using MDPs

The objective here is to learn the task of active localization in a multiple robot scenario. The MDP framework is used to effectively generalize the learning across different number of robots. Once it is trained for an appropriate number of robots required to capture the basic features in the map, the system would work for large range of number of robots without re-training.

Related Publications:

Multi-Robot Localization

Localization is one of the fundamental problems in mobile robotics. Localization is the process of determining the robot's position in the environment. In this research we strive to see how multiple robots can aid in localization of one another. We envisage a multi robotic scenario where several robots are in ambiguity about their states and require help of other robots to overcome their ambiguity.

Related Publications:

Optimal target tracking by multi sensor surveillance system

The objective is to see what kind of coordination mechanisms between sensors enables optimal detection of targets moving across a surveillance area.

Related Publications:

Cooperative collision avoidance

In an area that is prone to be criss crossed by several robots collision avoidance amongst them is inevitable. Here we see how cooperation between robots could enhance collision avoidance maneuvers between them.

Related Publications: