Nonparametric Continuous Sensor Registration
We develop a new mathematical framework that enables nonparametric joint semantic/appearance and geometric representation of continuous functions using data. The joint semantic and geometric embedding is modeled by representing the processes in a reproducing kernel Hilbert space. The framework allows the functions to be defined on arbitrary smooth manifolds where the action of a Lie group is used to align them. The continuous functions allow the registration to be independent of a specific signal resolution and the framework is fully analytical with a closed-form derivation of the Riemannian gradient and Hessian. We study a more specialized but widely used case where the Lie group acts on functions isometrically. We solve the problem by maximizing the inner product between two functions defined over data, while the continuous action of the rigid body motion Lie group is captured through the integration of the flow in the corresponding Lie algebra. Low-dimensional cases are derived with numerical examples to show the generality of the proposed framework. The high-dimensional derivation for the special Euclidean group acting on the Euclidean space showcases the point cloud registration and bird's-eye view map registration abilities. A specific derivation and implementation of this framework for RGB-D cameras outperform the state-of-the-art robust visual odometry and performs well in texture and structure-scares environments.
Robotic Exploration, Information Gathering, and Environmental Monitoring
The Incrementally-exploring Information Gathering (IIG) algorithms are sampling-based motion planning algorithms equipped with an information-theoretic convergence criterion for incremental informative motion planning. The proposed approach allows for dense map representation and incorporates the full state uncertainty into the planning process. The problem is formulated as a maximization problem with a budget constraint. Our approach is built on Rapidly-exploring Information Gathering (RIG) algorithms and benefits from the advantages of sampling-based optimal motion planning algorithms. Code: https://github.com/MaaniGhaffari/sampling_based_planners
A MATLAB implementation of the incremental Gaussian processes occupancy mapping using range-finder sensors is available here: https://github.com/MaaniGhaffari/incremental_gpom
Legged Robot State Estimation
We developed a state estimator, i.e, Right Invariant Extended Kalman Filter (RI-EKF) for legged robots that is robust to initialization and disturbance errors. As a result, it is suitable for robots that work in rough and unstructured environments. In particular, legged robots that are likely to engage in dynamic locomotion, running, and aggressive maneuvers benefit from this state estimator. Code: https://github.com/RossHartley/invariant-ekf
We also developed a long-term state estimation and mapping framework for legged robots that allows for reducing the drift and correcting the past estimates as the robot perceives new information. During long-term missions, odometry systems can drift substantially, leading to an unbounded growth in the covariance of the estimate and an undesirable expansion of the search space for data association tasks. We used the factor graph smoothing framework for building such systems in which real-time performance is achieved by exploiting the sparse structure of the Simultaneous Localization and Mapping (SLAM) problem.
Semantic Robotic Perception
Building on geometric-based techniques, semantic perception of the environment, such as outdoor terrain classes and indoor object categories can enable safe and efficient navigation as well as scene understanding in such environments. However, the incorporation of semantic knowledge into the map of the robot’s surrounding makes the map inference problem theoretically and computationally much more challenging. We tackle the problem of semantic sensor registration and real-time dense semantic robotic mapping for both outdoor and indoor environments to enable semantically-aware autonomous navigation.
The software for the work on Semantic ICP which is published in BMVC is available as an open source implementation at https://bitbucket.org/saparkison/semantic-icp
Radio Signal-based Localization and Tracking
Wireless Local Area Network (WLAN) and Bluetooth Low Energy (BLE) technologies are widespread and ubiquitous. Radio signals can be a complementary source of information alongside widely used sensors such as cameras. However, using cameras can raise privacy concerns, depending on the deployed location, which can limit the application of such systems. On the other hand, in indoors, radio signals are severely impacted due to shadowing and multipathing effects which make the available wireless-based positioning systems less accurate (1-10 m). We propose a radio-inertial localization and tracking system that exploits BLE, Inertial Measurement Unit (IMU), and magnetometer sensors with the quality available in standard smartphones.