Motion blur is a common problem for machine vision applications in legged mobile robotic platforms. Due to the oscillatory walking nature of these platforms, the cameras on them experience disturbances so that most of the captured frames are motion blurred. Motion blur results in loss of information in individual image frames and therefore makes single‐frame deblurring an ill‐posed problem. The variation in the magnitude of motion blur in consecutive video frames can be exploited to better restore these frames by combining the information available in them. The authors focus on the multi‐frame image deblurring problem and propose a promising method to achieve this information fusion especially when the video is corrupted by non‐linear blur, such as the case in legged mobile robots. The authors present the theory and results obtained on a realistic motion‐blur dataset collected on a representative robotic platform. These results are also compared with a leading single‐frame deblurring method and show improvement in particular for video corrupted by linear and non‐linear motion blur.

Sampling based methods resulted in feasible and effective motion planning algorithms for high dimensional configuration spaces and complex environments. A vast majority of such algorithms as well as their application rely on generating a set of open-loop trajectories first, which are then tracked by feedback control policies. However, controlling a dynamic robot to follow the planned path, while respecting the spatial constraints originating from the obstacles is still a challenging problem. There are some studies which combine statistical sampling techniques and feedback control methods which address this challenge using different approaches. From the feedback control theory perspective, Reference Governors proved to be a useful framework for constraint enforcement. Very recently, Arslan and Koditschek (2017) introduced a feedback motion planner that utilizes Reference Governors that provably solves the motion planning problem in simplified spherical worlds. In this context, here we propose a “trajectory-free” novel feedback motion planning algorithm which combines the two ideas: random trees and reference governors. Random tree part of the algorithm generates a collision-free region as a set of connected simple polygonal regions. Then, reference governor part navigates the dynamic robot from one region to the adjacent region in the tree structure, ensuring it stays inside the current region and asymptotically reaches to the connected region. Eventually, our algorithm robustly routes the robot from the start location to the goal location without collision. We demonstrate the validity and feasibility of the algorithm on simulation studies.

Mobile robots suffer from sensory data corruption due to body oscillations and disturbances. In particular, information loss on images captured with onboard cameras can be very high, and such loss may become irreversible or computationally costly to undo. In this paper, we propose a novel method to minimize average motion blur captured by such mobile visual sensors. To this end, we derive a motion blur metric (MMBM) that can be computed in real-time by using only inertial sensor measurements and validate it through comparisons with optic flow computations. The applicability of MMBM is illustrated through a motion blur minimizing system implemented on the SensoRHex hexapod robot by externally triggering an onboard camera based on MMBM values computed in real-time while the robot is walking straight on a flat surface. The resulting motion blur is compared to motion blur levels obtained with a regular, fixed frame-rate image acquisition schedule by both qualitative inspection and using a blind blur metric on captured images. MMBM based motion blur minimization system not only reduces average motion blur, but also avoids frames with extreme motion blur before an image gets corrupted by appropriately delaying the triggering of frame acquisition.

Pronking (aka. stotting) is a gait in which all legs are used in synchrony, resulting in long flight phases and large jumping heights that may potentially be useful for mobile robots on rough terrain. Robotic instantiations of this gait suffer from severe pitch instability either due to underactuation, or the lack of sufficient feedback. Nevertheless, the dynamic nature of this gait suggests that the Spring-Loaded Inverted Pendulum Model (SLIP), a very successful predictive model for both natural and robotic runners, would be a good basis for more robust and maneuverable robotic pronking. In this paper, we describe how “template-based control”, a controller structure based on the embedding of a simple dynamical “template” within a more complex “anchor” system, can be used to achieve stable and controllable pronking for a planar, underactuated hexapod model. In this context, high-level control of the gait is regulated through speed and height commands to the SLIP template, while the embedding controller based on approximate inverse-dynamics and carefully designed passive dynamics ensures the stability of the remaining degrees of freedom. We show through extensive simulation experiments that unlike existing open-loop alternatives, the resulting control structure provides stability, explicit maneuverability and significant robustness against sensor and actuator noise.

In this paper, we study the kinematics of a legged robot with half-circular leg morphology. In particular, our focus is on the RHex hexapod platform. A new kinematic model for RHex is developed considering the leg shape and its consequences, which was over simplified in the previous models seen in literature. The formulation is an accurate kinematic representation of the robot in the sagittal plane that is based on a four-link mechanism analogy. When only pure rolling motion of the legs are considered, it is found that when front and rear pairs of legs are in contact with the ground, the robot becomes a one degree-of-freedom mechanism and position of the middle pair of legs are redundant. The problem is solved in two steps; the first one being the determination of the initial configuration of the leg angular positions which defines the initial value of the variable distance of the front and rear leg and ground contact points. After the initial configuration of the system is set, pitch angle of the robot body can be manipulated by controlling one of the leg angular positions and the results are presented on an example case; positioning a body fixed unactuated sensor by controlling robot body pitch angle through the actuation of one of the legs. The results are a good display of the multi-functional aspect of the legs in addition to their use for locomotion.

This study presents a theoretical investigation of the rank-based multiple classifier decision problem for closed-set pattern classification. The case with classifier raw outputs in the form of candidate class rankings is considered and formulated as a discrete optimization problem with the objective function being the total probability of correct decision. The problem has a global optimum solution but is of prohibitive dimensionality. We present a partitioning formalism under which this dimensionality can be reduced by incorporating our prior knowledge about the problem domain and the structure of the training data. The formalism can effectively explain a number of rank-based combination approaches successfully used in the literature, one of which is discussed.

In this study, closed set, text-independent speaker identification is considered and the problem of improving the reliability of the decisions made by available algorithms is addressed. The work presented is based on the idea of combining the evidence from different algorithms or decision strategies to improve the recognition performance and the reliability. For this purpose, the models generated by a single algorithm for 17 speakers from the SPIDRE database are considered and a matrix of speaker-to-model fitness values is processed by two different decision strategies. Ideas from the mathematical theory of evidence are applied to combine the decisions produced by these two strategies to generate a better decision on the speaker identity. The combined decision show an improved degree of correctness hence suggesting a promising way of combining the decisions from partially successful algorithms.