Yeonsik Kang

Linear Tracking for a Fixed-Wing UAV Using Nonlinear Model Predictive Control

In this paper, a nonlinear model predictive control (NMPC) is used to design a high-level controller for a fixed-wing unmanned aerial vehicle (UAV). Given the kinematic model of the UAV dynamics, which is used as a model of the UAV with low-level autopilot avionics, the control objective of the NMPC is determined to track a desired line. After the error dynamics are derived, the problem of tracking a desired line is transformed into a problem of regulating the error from the desired line. A stability analysis follows to provide the conditions that can assure the closed-loop stability of the designed high-level NMPC. Furthermore, the control objective is extended to track adjoined multiple line segments. The simulation results demonstrate that the UAV controlled by the NMPC converged rapidly with a small overshoot. The performance of the NMPC was also verified through realistic ¿hardware in the loop simulation.¿

A Lidar-Based Decision-Making Method for Road Boundary Detection Using Multiple Kalman Filters

In this paper, a novel decision-making method is proposed for autonomous mobile robot navigation in an urban area where global positioning system (GPS) measurements are unreliable. The proposed method uses lidar measurements of the roads surface to detect road boundaries. An interacting multiple model method is proposed to determine the existence of a curb based on a probability threshold and to accurately estimate the roadside curb position. The decision outcome is used to determine the source of references suitable for reliable and seamless navigation. The performance of the decision-making algorithm is verified through extensive experiments with a mobile robot autonomously navigating through campus roads with several intersections and unreliable GPS measurements. Our experimental results demonstrate the reliability and good tracking performance of the proposed algorithm for autonomous urban navigation.

Dependable Humanoid Navigation System Based on Bipedal Locomotion

In this paper, a dependable humanoid navigation system is proposed by considering many difficulties in humanoid navigation based on bipedal locomotion in an uncertain environment. In particular, we propose a layered architecture to resolve complicated problems through a hierarchical manner. Within the proposed software architecture, a walking path planner, a walking footstep planner, and a walking pattern generator are integrated in a hierarchy to create a reliable motion that overcomes foot slippage and localization sensor noise. Each layer is designed to overcome difficulties originating from bipedal locomotion such as unstable dynamics, inclusion of a sinusoidal noise component in the localization sensor measurement, and disturbance regarding discrete footstepping. The designed navigation system is implemented on a human-sized experimental humanoid platform and is tested for the evaluation of its reliability and robustness in various tasks.

Emergency collision avoidance maneuver based on nonlinear model predictive control

In this study, nonlinear model predictive control (NMPC) is proposed for performing emergency collision avoidance maneuvers. NMPC is employed as a high-level controller that simultaneously controls the longitudinal and the lateral vehicle motion. The designed emergency collision avoidance controller considers the constraints of maximum wheel steering angle and maximum acceleration. NMPC predicts vehicle position using open-loop dynamics and calculates the optimized wheel steering control input and acceleration control input. A collision avoidance maneuver that does not consider vehicle dynamics may not avoid obstacles. In the worst case, it may transition into another dangerous situation. The performance of the proposed collision avoidance maneuver is simulated in Matlab/Simulink and CarSim, a realistic commercial simulation software.

Humanoid Path Planning From HRI Perspective: A Scalable Approach via Waypoints With a Time Index

This paper proposes a path planner for a humanoid robot to enhance its performance in terms of the human-robot interaction perspective. From the human point of view, the proposed method uses the time index that can generate a path that humans feel to be natural. In terms of the robot, the proposed method yields a waypoint-based path, the simplicity of which enables accurate tracking even for humanoid robots with complex dynamics. From an environmental perspective through which interactions occur, the proposed method can be easily expanded to a wide area. Overall, the proposed method can be described as a scalable path planner via waypoints with a time index for humanoid robots. Experiments have been conducted in test beds where the robot encounters unexpected exceptional situations. Throughout these trials, the robot successfully reached the goal location while iteratively replanning the path.

SOFTWARE ARCHITECTURE AND TASK DEFINITION OF A MULTIPLE HUMANOID COOPERATIVE CONTROL SYSTEM

This paper presents a cooperative control software architecture that coordinates a team of multiple humanoid to complete a mission by collaborating with each other. The mission of the humanoid team is decomposed into tasks and distributed to each humanoid to be executed. Each task is described by the proposed humanoid action primitives, which are designed to abstract broad classes of humanoid tasks appropriately. In particular, missions and tasks for the humanoid team are designed by using a finite state machine with a developed user interface. The multiple humanoid cooperative control software consists of 3 layers: the mission layer, task layer, and action layer. The software architecture has scalability to the number of humanoids and the number of assigned missions with its framework based on the CORBA middleware, which integrates many different functionalities of the humanoid. The feasibility and robustness of the implemented software architecture are verified through successful completion of the mission assigned to the humanoid team while each humanoid performs its given task sequentially.

Lane-level localization based on situational awareness on highway

This paper presents a lane-level localization algorithm based on situational awareness fusing radar, vision, GPS, and digital map on highway. Under the assumption that the maximum number of lanes is known by both GPS and digital map, the position of a vehicle is determined in a term of lane-level position via situational awareness on highway. The situational awareness aims at recognizing maneuver of the vehicle, guardrail, and surrounding vehicles. By fusion radar and vision sensors, it is proposed that the lane-level position can be determined inferring all outputs of situational awareness. The proposed algorithm is validated experimentally via field test data.

Probabilistic Mapping For UAV using Point-mass Target Detection

In this paper, a new method of building a probabilistic occupancy map for an unmanned aerial vehicle (UAV) equipped with a laser scanning sensor is proposed. For a UAV, target-tracking as well as mapping of obstacles are both important. Instead of using raw measurements to build a map, the proposed algorithm uses a well-known Interacting Multiple Model (IMM) based target formulation and tracking method to flrst process the noisy measurement data. The outputs from this process are used to recursively build the probabilistic occupancy map. The beneflt of this algorithm is attaining high quality occupancy map in spite of noisy sensor measurements simultaneously with the multiple-target tracking. In simulation, the obtained probabilistic occupancy map shows good agreement with the physical layout of the obstacles in the fleld. This shows the potential that the developed method can be used to help the unmanned vehicle navigate the fleld without prior knowledge. For an unmanned aerial vehicle (UAV), it is essential to be able to identify a safe path and localize its own position to perform a mission. Also, in order to perform a collision avoidance maneuver, the tracking of moving targets is necessary. A forward looking sensor such as a radar or laser scanner can help a UAV detect moving or stationary obstacles, especially when ∞ying in an outdoor environment. There are many tracking algorithms based on the Kalman fllter specialized in tracking a moving target. However, it is also necessary to produce a two-dimensional map of stationary obstacles to plan a safe path for the UAV. The method proposed in this thesis can achieve the tracking of both moving and stationary obstacles in a computationally e‐cient manner. The proposed method uses the outputs from the point mass target-tracking algorithm to generate a map of stationary obstacles. Therefore, the complex data association of large measurements is solved only once. In this paper, a new approach to build the probabilistic occupancy grid map based on the existing target-tracking algorithm will be presented assuming that the position of the vehicle is known. Probabilistic occupancy grid mapping has been very popular for robotic navigation applications and many papers have dealt with methods to update the known or unknown map and localize the position of the robot based on the measurements from sensors such as sonar or laser scanners. 1{3 The map update law has been extended to fllter out the measurements associated with dynamic obstacles in Ref. 4,5. Recently, the probabilistic occupancy grid map was constructed using a quadtree based approach which facilitates access to the information in the map and manages the memory e‐ciently. 6,7 In this paper, the Kalman fllter based Interacting Multiple Model (IMM) technique is used to employ a measurement model and to recursively build the probabilistic map of the sensor fleld. The localization of the unmanned vehicle is not addressed in this paper since our research goal is to provide a fast and robust

Performance Analysis of an IMM-Based Obstacle Detection Algorithm

In this paper, two approaches for obstacle detection and position estimation are presented. One is an algorithm based on M out of N detection logic and the other is an algorithm based on the probabilistic Interacting Multiple Model (IMM) method. The M out of N threshold-based algorithm declares that there is an obstacle present if it gets M validated measurements out of N consecutive measurements. IMM algorithm runs two different models in parallel, each based on a different hypothesis. One model assumes that there is an obstacle present while the other model assumes that there is no obstacle present in the sensor field of view. The performances of the two algorithms are compared based on their false alarm rate and detection speed. At first, Monte Carlo simulations are performed using only the false measurements to determine the thresholds for each method that generate a similar number of false detections. Using these thresholds, the detection speed of each method is compared and it is shown that the IMM-based algorithm is superior to the M out of N logic-based algorithm.Copyright

Real-time Obstacle Map Building with Target Tracking

In this paper, a new method is proposed to build a probabilistic occupancy map for an unmanned aerial vehicle (UAV) equipped with a forward-looking sensor, such as a laser scanning sensor (known as lidar). For a UAV, target tracking as well as mapping of obstacles are both important. Instead of using raw measurements to build a map, the proposed algorithm uses the interacting multiple model (IMM)-based target formulation and tracking method first to process the noisy measurement data. The state estimates and true target probability of each point-mass target tracks are then used to build a probabilistic occupancy map. Therefore, simultaneous tracking and mapping of both moving and stationary obstacles are accomplished in real time. In addition, the mapping algorithm has the robustness to the noisy sensor measurements. The obtained probabilistic occupancy map shows good agreement with the physical layout of the obstacles in the field in simulations. This shows the potential that the developed method can be used to help an unmanned vehicle navigate the field without a previous database of obstacles.