Thomas Engelhardt

Coding of arbitrarily shaped image segments based on a generalized orthogonal transform

Abstract Region oriented image representation offers several advantages over block-oriented schemes, e.g. adaptation to the local image characteristics, or object motion compensation as opposed to block-wise motion compensation. For the task of image data compression, i.e. image coding, new algorithms are needed which work on arbitrarily shaped image regions, called segments, instead of rectangular image blocks. Based on a generalized moment approach, the luminance function inside the segment is approximated by a weighted sum of basis functions, for example polynomials. A set of basis functions which is orthogonal with respect to the shape of the segment to be coded can be obtained using orthogonalization schemes. This results in the derivation of a generalized shape-adapted transform coder. Suitable coder and decoder structures are introduced which do not necessitate the transmission of the basis functions for each segment. Finally an application of the derived algorithms to image sequence coding at low data rates is shown, which is based on a segmentation of the motion compensated prediction error image.

Flatness-based control for a quadrotor camera helicopter using model predictive control trajectory generation

This paper addresses a flatness-based controller and a Model Predictive Control (MPC) trajectory generation for a quadrotor camera helicopter. Applications like aerial videography can highly benefit from an automation of the pilots tasks, enabling the camera operator to solely focus on camera motion control. The coupled nonlinear system dynamics of a quadrotor pose difficulties precisely controlling several channels simultaneously for agile maneuvering using conventional controllers. A flatness-based approach is employed to obtain linear input-output dynamics, even for large attitude angles. The associated state feedback equations are explicitly derived. The resulting linear system dynamics are controlled using a cascaded proportional control structure. Feasible reference trajectories are generated using a linear MPC, which translates operator commands for camera motions - e.g. relative to a point-of-interest - into quadrotor trajectories complying with operational constraints. Flatness-based controller and MPC trajectory generation show tracking errors below 1% in simulation tests. Accurate and smooth positioning is achieved in first indoor test flights. The gained results motivate adoptions of the proposed control approach to other UAV applications with similar demands for pilot automation and accuracy.

Wheelslide and Wheelskid Protection for a Single-Wheel Drive and Brake Module (SDBM) for Rail Vehicles

In this paper, the dynamic behaviour of a rail vehicle with single-wheel drive and brake modules is analysed. It is shown that in the wheelslide or wheelskid case, the linearised plant obtains an unstable pole whose location is determined by the shape of the current creep force curve. By considering possible variations in the shape of the creep force curve, the corresponding variation of the pole location in the right-half plane is calculated and a controller structure for creepage control and creep velocity control is suggested. If the controller design takes the worst-case pole location into account, the resulting controller stabilises the closed loop even at small velocities. Finally, this fact is illustrated through simulation results.

Multicopter unmanned aerial vehicle for automated inspection of wind turbines

A concept for a multicopter unmanned aerial vehicle (UAV) automatically performing inspection flights at a wind turbine is proposed. Key aspects of the concept are (1) a priori 3D mapping of the plant and (2) spline-based flight path planning as well as (3) a collision avoidance and distance control system. A quadrotor UAV prototype and its dynamical model are presented. Validation of the different aspects is carried out in simulation and partially in indoor tests using Robot Operating System (ROS). Existence of a 3D map is an essential precondition for path planning and collision-free flight. A brief initial flight preceding the actual inspection with a 2D LiDAR sensor yields a point cloud of the plant which is used for 3D mapping. This map is efficiently generated and represented using octrees, a hierarchical data structure that can be used for 3D maps. Subsequently a smooth and collision-free flight path is generated using splines. For redundancys sake navigation tasks not only rely on GPS but also on the LiDAR sensor mentioned before. The sensor allows for continuous estimation of the distance between multicopter and wind turbine. A distance control algorithm guarantees collision-free flight.

Propeller thrust identification and calibration for high-precision control of a quadrotor unmanned aerial vehicle

Quadrotors are becoming a major platform for both industrial and research activities, increasing the requirements for precise and dynamic control. Best control can be achieved by directly controlling the individual rotor thrust forces, which, during flight, can only be estimated. An identification of the thrust generation mechanism, and subsequent exploitation of this relation in the quadrotors command chain is therefore crucial. In this work, the relationship between a given angular velocity command and actual thrust of each individual rotor is modeled as a Wiener model, consisting of a linear dynamic first-order plus time delay element and a static nonlinear second order polynomial. Parameters are identified experimentally using a propeller thrust scale and measurements of the true rotor speed. In addition, an inflight calibration procedure is proposed for compensation of varying system and environment parameters. The command structure is then implemented on a quadrotor employed with a flatness-based controller, and indoor flights in a tracking cell show a low mean error of 0.2 m/s2 between commanded and true acceleration. The simplicity and accuracy of this approach motivates the application to quadrotors with all kinds of controllers to increase their performance during dynamic maneuvers.