Multi Robot Exploration and Mapping with Collision Avoidance

The objective is to obtain a topological representation of the environment incrementally built by a team of robots. A major challenge in an autonomous exploration task is to declare completion, which signifies that the entire environment is covered. Exploration and map building using multiple robots also need to deal with issues such as redundant exploration, collision avoidance, coordination and information exchange between the robots.

Related Publications:

    Multi-Robot Graph Exploration and Map Building with Collision Avoidance: A Decentralized Approach  [doi],
    Nagavarapu, S.C., Vachhani, L. and Sinha, A.
    Journal of Intelligent & Robotic Systems, vol. 83, no. 3, pp. 503-523, 2016.

    Generalizing Multi-Agent Graph Exploration Techniques,
    Nagavarapu, S.C., Vachhani, L., Sinha, A. and Buruily, S.
    Accepted, International Journal of Control, Automation, and Systems.

    A Decentralized Approach for Autonomous Multi-Robot Exploration and Map Building for Tree Structures  [doi],
    Nagavarapu, S.C., Vachhani, L. and Sinha, A.
    Proceedings of the Indian Control Conference (ICC), Chennai, India, pp. 274-279, 2015.

    A Robust Technique for Graph Exploration using Multiple Robots,
    Nagavarapu, S.C., Vachhani, L. and Sinha, A.
    (Manuscript in preparation).

    A Beam Steering based Framework for Continuous Target Tracking in Multi-Robot Systems,
    Nagavarapu, S.C., Vachhani, L. and Sinha, A.
    (Manuscript in preparation).




Vehicle Routing Algorithms and Travel Time Prediction for a Fleet of Autonomous (Driverless) Vehicles

The objective of this work is to design and develop routing algorithms for fleets of autonomous vehicles to perform mobility on demand services, scheduled vehicle transportation, etc. The system and the algorithms should be able to handle the varying traffic conditions, weather conditions, stochastic customer demands, etc. The FMS algorithms must be capable enough to handle such fluctuations in the traffic conditions, and re-route the vehicles accordingly. We also plan to extend the FMS algorithms to larger road networks. This work also aims to develop a high fidelity simulation engine to simulate the autonomous vehicle routing behavior.

Related Publications:

    Disruption Management for Dial-A-Ride Systems,
    Pandi, R.R., Ho, S.G., Nagavarapu, S.C., Tripathy, T. and Dauwels, J.
    Accepted, IEEE Intelligent Transportation Systems Magazine.

    A Generic GPU-Accelerated Framework for the Dial-A-Ride Problem,
    Pandi, R.R., Ho, S.G., Nagavarapu, S.C. and Dauwels, J.
    (Manuscript under revision).

    Solving Dial-A-Ride Problems using Multiple Ant Colony System with Fleet Size Minimisation  [doi],
    Tripathy, T., Nagavarapu, S.C., Azizian, K., Pandi, R.R. and Dauwels, J.
    Advances in Computational Intelligence Systems (ACIS), pp. 325-336, vol. 650, 2018.

    Development of a Simulation Platform to Implement Vehicle Routing Algorithms for Large Scale Fleet Management Systems  [doi],
    Nagavarapu, S.C., Tripathy, T. and Dauwels, J.
    Proceedings of the 20th IEEE International Conference on Intelligent Transportation Systems (ITSC): Workshop, Yokohama, Japan, pp. 124-129, 2017.

    Improved Tabu Search Heuristic for Static Dial-a-ride Problem: Faster and Better Convergence  [doi],
    Ho, S.G., Nagavarapu, S.C., Pandi, R.R. and Dauwels, J.
    Presented at the 25th Intelligent Transportation Systems World Congress (ITS-WC), Copenhagen, Denmark, 2018.

    Multi-atomic Annealing Heuristic for the Dial-a-ride Problem [doi],
    Ho, S.G., Pandi, R.R., Nagavarapu, S.C. and Dauwels, J.
    Proceedings of the 12th IEEE International Conference on Service Operations and Logistics, and Informatics (SOLI), Singapore, pp. 272-277, 2018.

    GPU-Accelerated Tabu Search Algorithm for Dial-A-Ride Problem  [doi],
    Pandi, R.R., Ho, S.G., Nagavarapu, S.C., Tripathy, T. and Dauwels, J.
    Proceedings of the 21st IEEE International Conference on Intelligent Transportation Systems (ITSC), Hawaii, USA, pp. 2519-2524, 2018.

    Solving Time-Dependent Dial-A-Ride Problem using Greedy Ant Colony Optimization  [doi],
    Ho, S.G., Koh, H.W., Pandi, R.R., Nagavarapu, S.C. and Dauwels, J.
    Proceedings of the 21st IEEE International Conference on Intelligent Transportation Systems (ITSC), Hawaii, USA, pp. 778-785, 2018.

    Deterministic Annealing for Depot Optimization: Applications to the Dial-A-Ride Problem  [doi],
    Pandi, R.R., Ho, S.G., Nagavarapu, S.C. and Dauwels, J.
    Proceedings of the 9th IEEE Symposium Series on Computational Intelligence (SSCI), Bangalore, India, pp. 88-95, 2018.

    Adaptive Algorithm for Dial-A-Ride Problem with Vehicle Breakdown  [doi],
    Pandi, R.R., Ho, S.G., Nagavarapu, S.C. and Dauwels, J.
    Proceedings of the 18th European Control Conference (ECC), Naples, Italy, pp. 2682-2688, 2019.

    Data-driven Models for Short-term Travel Time Prediction  [doi],
    Narayanan, A.K., Pranesh, C., Nagavarapu, S.C., Kumar, B.A. and Dauwels, J.
    Proceedings of the 22nd IEEE International Conference on Intelligent Transportation Systems (ITSC), Auckland, New Zealand, pp. 1941-1946, 2019.

    An Improved Tabu Search Heuristic for Static Dial-A-Ride Problem  [arXiv preprint],
    Ho, S.G., Nagavarapu, S.C., Pandi, R.R., Azizian, K. and Dauwels, J.
    (Manuscript under review).

    A Hybrid Tabu Search Algorithm for the Dial-A-Ride Problem with Fleet Size Minimization,
    Pandi, R.R., Ho, S.G., Nagavarapu, S.C., Tripathy, T. and Dauwels, J.
    (Manuscript under review).

    Lane Marker Detection and Rain Removal for Autonomous Vehicle Navigation,
    Nagavarapu, S.C., Li, S. and Dauwels, J.
    (Manuscript in preparation).

    Dynamic Object Removal in Point Clouds for Autonomous Navigation and Mapping,
    Muthuchamy, N., Nagavarapu, S.C. and Dauwels, J.
    (Manuscript in preparation).

Space Debris Removal in Low Earth Orbit (LEO) using Cubesats

The aim of this research proposal is to design a small satellite system to de-orbit orbital debris. Space debris in the near earth orbit is a growing concern for the space community. Recent studies on the instability of the debris population in the low Earth orbit (LEO, defined as the region below 2000 km altitude) indicate that the environment has reached a point where collisions among existing objects will force the LEO population to increase, at least in the next 200 years, even without any new launches. Since launches and LEO satellites are projected to grow exponentially in the coming decade, the debris problem in LEO is also expected to grow. Hence, clean-up of near Earth space from all the debris that has accumulated from defunct satellites and spent upper stages of launch vehicles will make it safer to put up satellites in orbit. In this study we intend to do a comprehensive system analysis study to understand what size and mass of debris can be deorbited from any given LEO altitude with a small physical device that can be the payload for a cubesat. The main objective is to perform the systems analysis study and finalize the most optimal deorbit payload and technique. This work also aims to develop machine learning algorithms that will accomplish the rendezvous and target acquisition.

Related Publications:

    Space Debris Removal from LEO using Cubesats: A System Level Study of Debris Capture & Deorbit Mechanisms and Debris Lifetime Analysis,
    Nagavarapu, S.C., Chandran, A., and Hastings, D.E.
    (Manuscript in preparation).

Optimal Fleet Size for Multi-Robot Exploration

The time to complete the exploration depends on the number of agents (team size or fleet size) used for exploration. Fleet size is often considered to be one of the important parameters in real time robotic applications such as surveillance, terrain mapping, search and rescue, etc. Finding an analytical expression for number of agents describing its relation with a given strategy, communication method and topology is challenging. The goal is to come up with a method to find the optimal number of agents for multi agent graph exploration.

Wheeler (Wireless Controlled Mobile Robot)

Wheeler is a wirelessly controllable mobile robot designed from scratch. The main objective is to construct a robot chassis from scratch, which is capable of moving to any direction, based on the command given using the wireless remote control. The robot is capable enough to carry smaller payloads (eg. laptop, small equipments, etc.). In order to have a light weight body, we have chosen steel chassis plates and wheels with rubber covering. PIC 16F877A microcontroller has ben used to program the robot motion controls.