This work focuses on developing collaborative exploration strategies that allow robots and humans to learn and co-adapt to exploit synergies within human-robot teams. Many existing collaborative strategies allocate tasks by leveraging each member’s strengths. By allowing robots to learn from their human teammates, humans can transfer some of their tasks to their robotic teammates enabling them to focus on tasks that are particularly challenging for robots, e.g., complex manipulation tasks. The goal of this work is to develop a framework that enables robots to learn domain specific knowledge to improve its execution of a collaborative human-robot tasks.

Initial results include a neural network-aided robot exploration strategy where domain expert knowledge, in conjunction with geometric and topological features around the exploration frontier, is used to improve the performance of a robot-only exploration strategy. The idea is for a robot to learn and apply domain expert knowledge during its selection of next best candidate location for exploration and mapping. We compared our approach to a baseline information-theoretic mapping and exploration strategy where a single robot is tasked to map and explore an unknown office-like environment. Our results shows how domain knowledge can improve the overall performance of the untrained strategy. In particular, our strategy shows a significant increase in the acquisition rate of semantic and topological information of the workspace during exploration.

This work is currently extended to enhance geology practice and workflow. We are collaborating with cognitive scientists and geologists to develop strategic data-collection routines and support scientific inferences in natural environments. Aerial robots equipped with onboard odometry, cameras, and lidars are particular well suited for these tasks since geologists are often interested in rock formations located in regions that are difficult to access by foot. Furthermore, collection of large amounts of surface and three dimensional data can be automated rather than rely on sparse manual measurements. Initial data collection efforts are focused in the Mecca Hills region in Southern California.

Being able to estimate and predict information about dynamic processes deepens our understanding of biological, chemical, and physical phenomena in the environment. Often, these dynamic processes exhibit complex, spatio-temporal behaviors. Mobile robots are particularly well-suited to monitor these processes because of their abilities to carry sensors and adapt their sensing locations. Robots can be used to support a wide range of activities dependent on tracking and predicting processes that vary across both space and time, such as tracking oil spills in water or pollutant concentrations in air for environmental monitoring, gas leaks for pipeline repair, or forest fire boundaries for search and rescue. For these processes, autonomous mobile robots modeling the environment and determining where to gather sensor measurements are cheaper than global tracking systems and more adaptive than fixed sensors. The process dynamics provide rich information about its spatial and temporal dependencies. Thus, robots should leverage their mobility and sensing capabilities to adequately model and estimate the environment.

However, given that they are inherently complicated, spatio- temporal processes are often difficult to model in a meaningful way, and even in scenarios where representations are available, they are often high-dimensional, which is computationally burdensome. Additionally, these processes often occur in dynamic, uncertain environments, so robots should not rely on centralized techniques to mitigate the effects of communication constraints and robot failures. The question then still remains as to how robots can leverage the spatio-temporal dynamics of the process to model and estimate the environment in a distributed way.

The goals of this project are to develop strategies to enable a team of mobile robots to adaptively sample and track a dynamic process. We create a distributed algorithm, where robots collect sparse sensor measurements, create a reduced-order model of a spatio-temporal process, and use this model to estimate field values for areas without sensor measurements of the dynamic process.

Videos

Teams of Four Marine Robots Tracking Dynamic Process

A concentration field was mapped and projected onto the tank using the video from a fluids experiment in an experimental flow tank. Four robots robots then tracked the projected concentration field using our developed distributed algorithm.

Papers

1. T. Salam, M.A. Hsieh. “Adaptive Sampling and Reduced Order Modeling of Dynamic Processes by Robot Teams,” IEEE Robotics and Automation Letters (RA-L), 4(2), April 2019, pp.  477-484 Bibtex | PDF

Optimal Paths in Flows

Autonomous marine vehicles (AMVs) are typically deployed for long periods of time in the ocean to monitor different physical, chemical, and biological processes. Given their limited energy budgets, it makes sense to consider motion plans that leverage the dynamics of the surrounding flow field so as to minimize energy usage for these vehicles.  This project focuses on developing suitable graph search based techniques to compute energy and/or time optimal paths for AMVs in two- and three-dimensional time-varying flows (2D+1 and 3D+1).   This project has contributed novel techniques that can capture the kinematic actuation constraints on the vehicles in our cost functions, generate optimal paths in different homotopy classes, and employ an adaptive discretization scheme to construct the search graph.  Our current efforts are focused on how best to leverage coherent structure information into our strategies.

This work is a collaboration between Dr. Subhrajit Bhattacharya at Lehigh University.

Papers

1. D. Kularatne, M. A. Hsieh, and E. Forgoston.  “Using Control to Shape Stochastic Escape and Switching Dynamics,” Chaos 29, 053128 (2019); https://doi.org/10.1063/1.5090113.  Bibtex | PDF
2. D. Kularatne, S. Bhattacharya, and M. A. Hsieh. “Going With The Flow: A Graph Based Approach to Optimal Path Planning in General Flows,” Autonomous Robots: RSS 2016 Special Issue, 42(7), Oct 2018, pp 1369 — 1387. Bibtex | PDF
3. D. Kularatne, E. Forgoston, and M. A. Hsieh. “Exploiting Stochasticity for Navigation in Gyre Flows,” in the Proc. of the 2018 Robotics: Science and Systems (RSS 2018), Jun 2018, Pittsburgh, PA USA. Bibtex | PDF
4. H. Hajieghrary, D. Kularatne, and M. A. Hsieh. “Differential Geometric Approach to Trajectory Planning: Cooperative Transport by a Team of Autonomous Marine Vehicles,” in the Proc. of the 2018 IEEE American Control Conference, Jun 2018, Milwaukee, WI USA. Bibtex | PDF
5. D. Kularatne, H. Hajieghrary, and M. A. Hsieh. “Optimal Path Planning in Time-Varying Flows with Forecasting Uncertainties,” in the Proc. of the 2018 IEEE International Conference on Robotics and Automation (ICRA2018), May 2018, Brisbane, Australia. Bibtex | PDF
6. D. Kularatne, S. Bhattacharya, and M. A. Hsieh. “Optimal Path Planning in Time-Varying Flows using Adaptive Discretization,” IEEE Robotics and Automation Letters (RA-L), 3(1), Jan 2018, pp. 458-465.  Bibtex | PDF
7. D. Kularatne and M. A. Hsieh. “Tracking Attracting Manifolds in Flows,” Autonomous Robots, 41(8), Mar 2017, pp. 1575–1588. Bibtex | PDF
8. D. Kularatne, S. Bhattacharya, M. A. Hsieh. “Time and Energy Optimal Path Planning on a Flow Field,” in the Proc. of the 2016 Robotics: Science and Systems (RSS2016), Jun 2016, Ann Arbor, MI USA. Bibtex | PDF

Path Planning for a Magnetic Millirobot

For actuation of small-scale robotic systems, magnetic control methods have garnered significant interest since magnetic fields can be selectively applied without affecting non-magnetic materials. Existing approaches for the control of single or multiple microrobots operating in magnetic fields have mostly focused on two aspects of the problem: 1) design of the physical geometry of the robot, and 2) design of devices to manipulate the local magnetic field. Instead of focusing on ways of manipulating the local magnetic field or the physical geometry of the robot, this project investigates how global design of the field topology can be leveraged for control and manipulation of microrobots. A significant advantage of leveraging topological features of the force vector field is that it does not require complete knowledge of the field and yet results in comparable performance to vehicles following optimal paths where the vector field has been explicitly accounted for.

We create a path planning and trajectory following strategy for a magnetic millirobot that leverages the nonlinearities in the external magnetic force field (MFF). The strategy creates a library of candidate MFFs and characterizes their topologies by identifying the unstable manifolds in the workspace. The path planning problem is then posed as a graph search problem where the computed path consists of a sequence of unstable manifolds segments and their associated MFFs. By tracking the robot’s position and sequentially applying the MFFs, the robot navigates along each unstable manifold until it reaches the goal.

Videos

Papers

1. A. Mansfield, D. Kularatne, E. Steager, and M. A. Hsieh, “A Topological Approach to Path Planning for a Magnetic Millirobot,” Submitted to Proc. of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS2020), Under Review, Oct 2020, Las Vegas, NV.

This work leverages the complimentary mobility and sensing capabilities of a network of heterogeneous robots that operate in remote oceanic environments, to efficiently collect information and effectively manage the volume of data collected. We consider the deployment of minimally actuated active drifters or similarly power-constrained mobile sensors. The active drifters periodically offload their sensory information to more capable robotic vehicles routed in a coordinated fashion through the drifter ensemble. Different from their passive counterparts, active drifters can adapt, albeit in a limited fashion, their sampling strategies to maximize information gain. When coupled with more capable autonomous surface, underwater, or even remotely operated vehicles (ASVs, AUVs, or ROVs), active drifters can significantly increase the spatial sampling reach of ASVs, AUVs, and ROVs. On the other hand, ASVs, AUVs, and ROVs can complement the sensing capabilities of active drifters since they have larger sensor payloads and reach regions not easily accessible to the active drifters due to actuation limitations. However, due to their severely limited power budgets, active drifters
must have the ability to plan and execute energy aware motion control and coordination strategies for data harvesting and rendezvous with AVUs and ASVs for data exchange and upload.

This work focuses on developing motion planning and control strategies for teams of mobile sensors with limited actuation capabilities or power budgets, i.e., active drifters, to harvest data and rendezvous with other autonomous vehicles. The proposed paradigm maximizes the impact of small, power constrained mobile sensors by leveraging the surrounding environmental dynamics to reduce their energy requirements.  The objectives of this work are:

1. To develop energy-aware motion control strategies for synchronous rendezvous by leveraging the dynamics of the fluidic environment;
2. To synthesize distributed synchronous rendezvous strategies for a set of moving rendezvous
   points whose motions are dictated by surrounding flow field;
3. To develop a stochastic modeling and control framework where the individual average behavior can be specified and tuned to achieve the desired collective targets; and
4. To validate the proposed strategies using the multi-robot Coherent Structure Testbed (mCoSTe).

Success of these endeavors will improve the autonomy and energy efficiency of various marine
platforms, directly affect the human’s abilities to navigate the oceans, increase the energy-efficiency of existing robotic sensor networks, and provide greater situational awareness for marine, coastal, and littoral applications.  The research focuses on developing a general stochastic control framework for coordinated energy-aware motion planning and navigation that are important for power constrained unmanned systems. The expected outcomes include:

• Energy aware motion planning and control strategies for minimally actuated autonomous vehicles with limited power budgets;
• New stochastic control tools to enable large collectives of autonomous sensing resources to rendezvous in dynamic environments; and
• A novel modeling, control, and analysis framework for coordinating large collectives of collaborating autonomous unmanned systems subject to energy constraints.

This is a collaborative effort with Dr. Herbert Tanner’s group at the University of Delaware.