Online Multi-modal Learning and Adaptive Informative Trajectory Planning for Autonomous Exploration

@inproceedings{Arora2017OnlineML,
  title={Online Multi-modal Learning and Adaptive Informative Trajectory Planning for Autonomous Exploration},
  author={Akash Arora and P. Michael Furlong and Robert C. Fitch and Terrence Fong and Salah Sukkarieh and Richard Elphic},
  booktitle={FSR},
  year={2017}
}
In robotic information gathering missions, scientists are typically interested in understanding variables which require proxy measurements from specialized sensor suites to estimate. However, energy and time constraints limit how often these sensors can be used in a mission. Robots are also equipped with cheaper to use navigation sensors such as cameras. In this paper, we explore a challenging planning problem in which a robot is required to learn about a scientific variable of interest in an… 
Multi-modal active perception for information gathering in science missions
TLDR
This paper presents an active perception approach that aims to reduce robots’ reliance on human supervision and improve science productivity by encoding scientists’ domain knowledge and decision making process on-board by presenting a Bayesian network architecture to compactly model critical aspects of scientific knowledge while remaining robust to observation and modeling uncertainty.
An Adaptive Memory-Based Reinforcement Learning Controller
TLDR
A reinforcement learning based controller to provide a universal solution for controlling autonomous robots in the field of exploration that includes a memory of past actions and rewards to efficiently analyze an agent’s current state when planning future actions.
Path Planning in Dynamic Environments using Generative RNNs and Monte Carlo Tree Search
TLDR
An integrated path planning framework using generative Recurrent Neural Networks within a Monte Carlo Tree Search (MCTS) using a learnt model of social response to predict crowd dynamics during planning across the action space is proposed.
Foraging, Prospecting, and Falsification - Improving Three Aspects of Autonomous Science
TLDR
The thesis proven in this work is that accounting for operational and environmental context improves science autonomy algorithms, and each of these algorithms improve components of autonomous science, and thereby the process as a whole.

References

SHOWING 1-10 OF 21 REFERENCES
Near-optimal Observation Selection using Submodular Functions
TLDR
Recent advances in systematically exploiting submodularity property to efficiently achieve near-optimal observation selections are surveyed and the effectiveness of these approaches is illustrated on problems of monitoring environmental phenomena and water distribution networks.
An approach to autonomous science by modeling geological knowledge in a Bayesian framework
TLDR
This paper presents an approach to extend the high level autonomy of robots by enabling them to model and reason about scientific knowledge on-board by using Bayesian networks to encode scientific knowledge and adapting Monte Carlo Tree Search techniques to reason about the network and plan informative sensing actions.
Efficient multi-sensor exploration using dependent observations and conditional mutual information
TLDR
A multimodal exploration and mapping approach that extends an occupancy grid map formulation to incorporate conditionally dependent sensor observations from multiple sensors and enables reasoning about uncertainty to select maximally informative actions is presented.
Modeling curiosity in a mobile robot for long-term autonomous exploration and monitoring
TLDR
A realtime topic modeling technique is used to build a semantic perception model of the environment that is suitable for long-term exploration missions and is able to do tasks such as coral reef inspection, diver following, and sea floor exploration without any prior training or preparation.
Sampling-based robotic information gathering algorithms
TLDR
This work proposes three sampling-based motion planning algorithms for generating informative mobile robot trajectories, and provides analysis of the asymptotic optimality of these algorithms, and presents several conservative pruning strategies for modular, submodular, and time-varying information objectives.
Science Autonomy for Rover Subsurface Exploration of the Atacama Desert
TLDR
In these field experiments the remote science team uses a novel control strategy that intersperses preplanned activities with autonomous decision making and performs automatic data collection, interpretation, and response at multiple spatial scales.
Active planning for underwater inspection and the benefit of adaptivity
TLDR
This work formulate the inspection planning problem as an extension to Bayesian active learning, and it proves that the potential benefit of adaptivity can be reduced from an exponential to a constant factor by changing the problem from cost minimization with a constraint on information gain to variance reduction with a constrained cost.
Hierarchical probabilistic regression for AUV-based adaptive sampling of marine phenomena
TLDR
This work uses organism abundance models along with spatial models of environmental features learned immediately after AUV deployments to compute spatial distributions of organisms in the coastal ocean purely from in situ AUV data.
Robots for Environmental Monitoring: Significant Advancements and Applications
TLDR
This article collates and discusses the significant advancements and applications of marine, terrestrial, and airborne robotic systems developed for environmental monitoring during the last two decades.
Coverage for robotics – A survey of recent results
  • H. Choset
  • Computer Science
    Annals of Mathematics and Artificial Intelligence
  • 2004
TLDR
This paper surveys recent results in coverage path planning, a new path planning approach that determines a path for a robot to pass over all points in its free space, and organizes the coverage algorithms into heuristic, approximate, partial-approximate and exact cellular decompositions.
...
1
2
3
...