A Novel Algorithm for Time Optimal Trajectory Planning

Time optimal trajectory planning under various hard constraints plays a significant role in simultaneously meeting the requirements on high productivity and high accuracy in the fields of both machining tools and robotics. In this paper, the problem of time optimal trajectory planning is first formulated. A novel back and forward check algorithm is subsequently proposed to solve the minimum time feed-rate optimization problem. The basic idea of the algorithm is to search the feasible solution in the specified interval using the back or forward operations. Four lemmas are presented to illustrate the calculating procedure of optimal solution and the feasibility of the proposed algorithm. Both the elliptic curve and eight profile are used as case studies to verify the effectiveness of the proposed algorithm.

Multi-Model Adaptive Regulation for Aircraft Dynamics With Different Zero Structures

An adaptive regulator is designed for parameter dependent families of systems subject to changes in the zero structure. Since continuous adaptive regulation is limited by relative degree and right half plane zeros, a multiple model adaptive regulator is implemented. The two multiple model design subproblems, covering and switching, are addressed with LQR state feedback and Lyapunov function switch logic respectively. These two subproblems are combined into a set of Linear Matrix Inequalities (LMI) and concurrently solved. The multiple model design method is applied to longitudinal aircraft dynamics.

A Review of Target Pursuit Strategies in Aerial Species

Aerial pursuit in nature is a complex task that involves interaction with targets in motion. To date, many researchers have analyzed aerial predation strategies used by different flying species for the pursuit and interception of targets such as a prey or a conspecific. In this article, we provide a brief review of these different predation strategies with the focus primarily on insects and bats that rely on different sensory variables (vision and sonar) for navigation. The Knowledge gained from studying these strategies can guide the development of bio-inspired approaches for navigation of engineered systems.

Design and Dynamic Modeling of a Flexible Feathering Joint for Robotic Fish Pectoral Fins

In this paper, we propose a novel design for a pectoral fin joint of a robotic fish. This joint uses a flexible part to enable the rowing pectoral fin to feather passively and thus reduce the hydrodynamic drag in the recovery stroke. On the other hand, a mechanical stopper allows the fin to maintain its motion prescribed by the servomotor in the power stroke. The design results in net thrust even when the fin is actuated symmetrically for the power and recovery strokes. A dynamic model for this joint and for a pectoral fin-actuated robotic fish involving such joints is presented. The pectoral fin is modeled as a rigid plate connected to the servo arm through a pair of torsional spring and damper that describes the flexible joint. The hydrodynamic force on the fin is evaluated with blade element theory, where all three components of the force are considered due to the feathering degree of freedom of the fin. Experimental results on robotic fish prototype are provided to support the effectiveness of the design and the presented dynamic model. We utilize three different joints (with different sizes and different flexible materials), produced with a multi-material 3D printer, and measure the feathering angles of the joints and the forward swimming velocities of the robotic fish. Good match between the model predictions and experimental data is achieved, and the advantage of the proposed flexible joint over a rigid joint, where the power and recovery strokes have to be actuated at different speeds to produce thrust, is demonstrated.

A Computationally Efficient Predictive Controller for Lane Keeping of Semi-Autonomous Vehicles

Model predictive control (MPC) is a popular technique for the development of active safety systems. However, its high computational cost prevents it from being implemented on lower-cost hardware. This paper presents a computationally efficient predictive controller for lane keeping assistance systems. The controller shares control with the driver, and applies a correction steering when there is a potential lane departure. Using the explicit feedback MPC, a multi-parametric nonlinear programming problem with a human-in-the-loop model and safety constraints is formulated. The cost function is chosen as the difference between the linear state feedback function to be determined and the resultant optimal control sequence of the MPC problem solved off-line given the current state. The piecewise linear feedback function is obtained by solving the parametric programming with an approximation approach. The effectiveness of the controller is evaluated through numerical simulations.

LMI Based Control of Stick-Slip Oscillations in Drilling

This paper investigates control of stick-slip oscillations in drilling from a linear matrix inequality perspective. Stick-slip oscillations include a period of no angular motion at the bit caused by a large static friction torque followed by a period of rapid angular motion at the bit caused by a build up of torque in the drilling pipe. Many of the model parameters are uncertain but belong to convex sets, and the friction torques are not easily modeled. The linear matrix inequality approach facilitates design of state feedback controllers in the presence of polytopic uncertainties and can be optimized to reject disturbance effects relative to outputs. Results indicate that the linear matrix inequality approach leads to a simple controller, successfully alleviates the stick-slip problem, and is robust to uncertainties.

A Discrete-Time Nonlinear Estimator for an Orthosis-Aided Gait

To achieve automatic operation of a powered orthosis-aided gait or functional electrical stimulation-based walking restoration, accurate estimation of the leg angles is of utmost importance. Various phases of walking last for a short duration of time; thus, an accurate estimator is required with a fast convergence rate. To overcome this challenge, this paper presents a discrete-time nonlinear estimation algorithm to estimate lower-limb angles during an orthosis-aided walking. To this end, we use measurements from 6 degree-of-freedom (DOF) inertial measurement units (IMUs) to estimate the lower limb angles. The estimator is based on a state-dependent coefficient (SDC) linearization or extended linearization of the nonlinear functions. A combination of multiple discrete SDCs is used to compute an optimal gain of the nonlinear estimator based on uncertainty minimization criteria. The nonlinear estimator is robust to uncertainties in system modeling and sensor noise/bias from the IMUs. Monte Carlo simulation studies reveal that the estimator outperforms widely used discrete-time extended Kalman (EKF) filter with respect to average root-mean squared estimation error (RMSE) criteria.

Autonomous Lighting Audits: Part 2 — Light Identification and Analysis

Buildings are responsible for approximately 40% of all US energy use and carbon emissions. There exists large potential to improve building efficiency through retro-commissioning, but expense and required expertise of building auditors limit current implementation. Autonomous robotic assessments have the potential to provide consistent building energy audits with reduced cost and enhanced capabilities. As a first step in automating building audits, this paper presents work on automating the lighting analysis of a building. As an aerial vehicle navigates and explores a room, the prototype system captures images and collects spectrometer readings. These data are used to quantify and classify lighting in a room. Additionally, images acquired from the optical camera are merged to form a composite image of the area. This composite image is used for navigation to lights to record spectrometer readings. Lighting type is then classified from these spectrometer readings. The combined lighting quantification and classification is used to create a topology map of light levels. The combined data are used to perform a thorough analysis of lighting and make lighting recommendations.

Bang-Bang Trajectory Plans With Dynamic Balance Constraints: Fast Rotational Reconfigurations for RoboSimian

This paper investigates the problem of rapidly transitioning the pose of a limbed robot while remaining balanced. In particular, we consider motions where rotational accelerations may significantly affect the center of pressure location within a limited base of support. We consider solutions for high-impedance robots with stiff, high-torque actuators that essentially provide accurate, position-control outputs at the joints. We present and compare three methods for generating joint trajectories to achieve fast yet feasible dynamic motions for such systems while maintaining a safety margin for the center of pressure location, toward robust balance. We focus on development of theory and intuition for each method and quantify performance in terms of achievable speed of transition and required joint velocity limits.

Stability Analysis of Nonlinear Connected Vehicle Systems

In this paper, we investigate the nonlinear dynamics of connected vehicle systems. Vehicle-to-vehicle (V2V) communication is exploited when controlling the longitudinal motion of a few vehicles in the traffic flow. In order to achieve the desired system-level behavior, the plant stability and the head-to-tail string stability are characterized at the nonlinear level using Lyapunov functions. A motif-based approach is utilized that allows modular design for large-scale vehicle networks. Stability analysis of motifs are summarized using stability diagrams, which are validated by numerical simulations.