Autonomous mobile robot model predictive control

2015 IEEE International Instrumentation and Measurement Technology Conference (I2MTC) Proceedings

Spinello

et al. 2015

Abstract-Among the major challenges in tracking a predefined trajectory of a nonholonomic mobile robot operating in indoor environments is to determine an appropriate feedback control. In the technical literature, numerous controllers have been proposed to date for solving trajectory tracking and/or regulation problems of mobile robots. Most of them are quite promising to implement and mobile robots can operate in quite complex environments with these controllers. Nevertheless, one of the shortcomings of most of the existing controllers is the requirement of sophisticated hardwares, which, in some cases, more costly than the mobile robot itself. In addition, the analytical expression for most of the controllers is quite complex even for a simple nonholonomic system, an unicycle, for example. In this paper, we propose a neghboring optimal controller coupled with the state estimation for a differential drive mobile robot to track a pre-defined trajectory using signal strength measurements from RF (radio frequency) sensors placed in its operating environment. We emphasize that the RF sensors have limited communication range with the robot. The proposed controller is easy to implement, does not require sophisticated hardware for processing, and has relatively easier expression for the feedback controller. A numerical example is provided to validate the theoretical results. Computer simulations will be backed by experimental results for future investigations.

Section: ∆Q(t) = F(t)∆q(t)+g(t)∆u(t)+l(t)ξ(t) ∆Q(0) = ∆Qmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Neighboring optimal control for mobile robot trajectory tracking with range-limited sensors

2015 IEEE International Instrumentation and Measurement Technology Conference (I2MTC) Proceedings

Spinello

et al. 2015

“…(a) initialize estimated pose and state error covariance matrix as in (34). (b) use (35) to compute error covariance matrices, W C and N C .…”

Section: S(t) =F(t)s(t) + S(t)fmentioning

confidence: 99%

“…For wheeled mobile robots, conventional control laws have been applied for solving tracking problems [58,30,32,43,1,23,49] and stabilization problems [3,17,51,54,8]. For example, see [29,28,39,48,12,14] for backstepping methods [11,24,53] for sliding mode control, [9,34,18] for moving horizon H ∞ tracking control coupled with disturbance effect, and [47] for transverse function approach. A vector-field orientation feedback control method for a differentially driven wheeled vehicle has been demonstrated in [46].…”

Section: Introductionmentioning

confidence: 99%

RFID-Based Mobile Robot Trajectory Tracking and Point Stabilization Through On-line Neighboring Optimal Control

2014

J Intell Robot Syst

In this manuscript, we propose an on-line trajectorytracking algorithm for nonholonomic Differential-Drive Mobile Robots (DDMRs) in the presence of possibly large parametric and measurement uncertainties. Most mobile robot tracking techniques that depend on reference RF beacons rely on approximating line-of-sight (LOS) distances between these beacons and the robot. The approximation of LOS is mostly performed using Received Signal Strength (RSS) measurements of signals propagating between the robot and RF beacons. However, accurate mapping between RSS measurements and LOS distance remains a significant challenge and is almost impossible to achieve in an indoor reverberant environment. This paper contributes to the development of a neighboring optimal control strategy where the two major control tasks, trajectory tracking and point stabilization, are solved and treated as a unified manner using RSS measurements emitted from Radio Frequency IDentification (RFID) tags. The proposed control scheme is divided into two cascaded phases. The first phase provides the robot's nominal control inputs (speeds) and its trajectory using full-state feedback. In the second phase, we design the neighboring optimal controller, where RSS measurements are used to better estimate the robot's pose by employing an optimal filter. Simulation and experimental results are presented to demonstrate the performance of the proposed optimal feedback controller for solving the stabilization and trajectory tracking problems using a DDMR. Keywords Mobile robot navigation · RFID systems · optimal control · trajectory tracking · robot stabilization · nonholonomic systems. Frequently Used Symbols K(t)Feedback control gain at time t H K Hamiltonian's gradient with respect to K s Number of RFID tags in the environment ψ Costate variable (Lagrange multiplier)Robot's actual and desired pose at time t q j t ∈ R 3 jth tag position in 3D space t 0 ,t f Initial and final time instantsRobot's control input vector at time t ξ (t) Robot's actuator noise at time t ζ (t)Measurement noise vector at time t (·) o , (·) ε , (·) ad Optimal, perturb, admissible value of (·) L Lebesgue measurable function space ν T dJ(·) Gateaux (directional) derivative of J in direction ν 1 Introduction

“…Numerous nonlinear control laws have been proposed in the literature to address the trajectory tracking problem of mobile robots, see [3], [4], [5] for backstepping methods, [6], [7], [8] for sliding mode control, [9], [10], [11] for moving horizon H ∞ tracking control coupled with disturbance effect, 1 http://en.wikipedia.org/wiki/Curiosity (rover) 2 http://en.wikipedia.org/wiki/Google driverless car 3 http://www.youtube.com/watch?v=YgEUrkY80-U 4 http://news.stanford.edu/news/2012/august/surfing-robot-082312.html [12] for transverse function approach, [13], [14] for formation control, and [15] for optimal motion planning. A vector-field orientation feedback control method for a differentially driven wheeled vehicle has been demonstrated in [16].…”

Section: Introductionmentioning

confidence: 99%

Linear time-invariant feedback operator for mobile robot trajectory tracking

2015 IEEE International Instrumentation and Measurement Technology Conference (I2MTC) Proceedings

Farkas

et al. 2015

Abstract-In this paper we propose a linear time-invariant (LTI) state feedback operator for nonholonomic mobile robots to track a pre-defined trajectory on a 2D planar region. Designing an on-line smooth state feedback control law is still among the major challenges for solving the trajectory tracking problem of nonholonomic systems including differential drive mobile robots. To address this problem, numerous nonlinear feedback control laws have been proposed to date. Most of these feedback control laws yield online solutions to the trajectory tracking problem with satisfactory tracking errors asymptotically. In most cases, they suffer from an overwhelming degree of computational complexity even to track a 2D trajectory using a simple unicyclelike nonholonomic system. The proposed control law offers an advantage of being smooth and LTI state-feedback in addition to its capability to address problems of partially observed systems. The shortcoming of the proposed control law is that the LTI feedback operator is computed off-line before the robot applies its control signal into the left and right wheels through their actuators. The theoretical results are supported by computer simulations followed by experiments with an e-puck mobile robot.