Study on Motion Tracking Control System for AGV

Download Full-Text PDF Cite this Publication

Text Only Version

Study on Motion Tracking Control System for AGV

Wallelgn Yonas Akele1

Tianjin Key Laboratory of Information Sensing and Intelligent Control

School of Automation and Electrical Engineering Tianjin University of Technology and Education, Tianjin 300222,P.R.China

Prof. Geng Huang-Yang2

Tianjin Key Laboratory of Information Sensing and Intelligent Control

School of Automation and Electrical Engineering Tianjin University of Technology and Education, Tianjin 300222,P.R.China

Abstract:-This paper presents an automatic guided vehicle (AGV) motion tracking control system. With the application of new control algorithms and the development of electronic technology, AGV is developing toward high speed, high precision, openness, intelligence, and networking, and it also puts forward higher requirements for motion control systems. To realize high-speed and high-precision position control and trajectory tracking, AGV must rely on advanced control strategies and excellent motion control systems. In this paper, according to the requirements of the control system, the Arduino/51/STM 32 microcontroller is selected as the core to design the motion control system. The paper studies the automatic guided vehicle motion tracking control system. The simulation of the motion tracking control of the intelligent robot is performed using MATLAB/Simulink. The Simulink model as well as the graph showed the AGV can reach the moving goal successfully.

Keywords: Automated guided vehicle (AGV), Motion tracking control system, obstacle avoidance, Simulink model, PID controller.

  1. INTRODUCTION

    An automated guided vehicle is a programmable mobile vehicle that follows marked lines or ground wires. Automatic steering vehicles are robots that run on the floor of a facility run by a combination of software and sensor management systems.

    The AGVs (Automated guided vehicle) began as transport devices developed to assist the manufacturing system. In the industrial robotic field, they are defined as transport vehicles driven by a computer system with different mechanical configurations [1]. Earlier inventions on AGVs can be dated back to Barrett Electronics in 1953. One of the oldest publications on AGV can be found in [2]. It is used to distribute materials in warehouses and to move and operate in production facilities to production areas and storage areas. It consists of different components like cranes and hoists, elevator and lifts, Conveyors, Robots, Automated Storage and Retrieval system (AS/RS)) and so on which are focused on the process of transferring something from one place to another. Utilization of components, either individually or from a combination point of view, is determined by its application or pre-assigned flexibility. Tracking is one of the most important aspects of AGV control and a prerequisite for completing accurate road tracking work. However, AGV is very indirect, making it difficult to track control challenging [3]. Many methods were proposed to solve this problem [4] and [5] proposed a linear proportional control method, [6]

    proposed a PID control method. This intelligent handling robot is based on the Arduino/51/STM32 microcontroller board program, to design an aluminum alloy body, multiple mounting holes on the robot chassis adding various sensors and controller, a space for the servo is also left to turn the mechanical manipulator. With the help of the engine, the tank chassis can turn smoothly to the left, right, circular, forward, and robot backward, etc. The size of the robot body part is 290mm length, 520mm height and 260mm width in thickness, where it adopts intelligent bus servos, and supports various controls, metal gears, and imported potentiometers are mounted on the arm of the robot and the robot has 6 degrees of freedom robotic arms with every joint controlled separately.

  2. OBSTACLE DETECTION

    In a real environment, an automated guided vehicle must avoid obstacles to go to a target. Depending on the positions of the target or the goal and the obstacle (s) relative to the automated guided vehicle, as the automated guided vehicle moves toward the target and sensors detect obstacles, it is important to control the avoiding strategy and speed. When there are barriers or obstacles in the environment, the robot's response is based on sensory information of the obstacles and the targeted position.

    In addition to this, sensors determine if something or an object was in the forward motion of the vehicle. They detect something; the robot will pause before checking if the obstacle is still present. For obvious reasons, the sensors are temporarily disabled as automated guided vehicle approaches to the wall the rear bumper sensors were activated to detect the rear collision detection. If this event occurred, the vehicle will be permanently disabled.

  3. HARDWARE IMPLEMENTATION USING DIFFERENT SENSORS

Ultrasonic sensor:- the ultrasonic range detection module can realize the non-contact ranging function of 2cm-300cm and has 40 kHz frequency specification. It requires a power supply voltage is 5V with a working current specification as 15mA. When the pin of the range is suspended, the range is 3m, The ultrasonic sensor can convert the measured distance into an analog voltage output, and the output voltage is directly proportional to the measured distance. The distance can be calculated with the following formula Distance=time*speed

Color/Gray sensor:-To distinguish colors according to the intensity of light reflection by different colors. It is characterized by simple detection and can be analyzed according to the collected AD value.

d

dt

Where r

(3)

and i are angular velocity of the right sprocket

Singletrack sensor:- The infrared emitting diode of the sensor continuously emits infrared rays. When the emitted infrared rays aren't reflected or reflected but aren't strong enough. The infrared receiving tube is always in the off state, At this time, the output end of the module is high level and the indicating diode is always in the output state. When the detected object appears within the detection range, the infrared ray is reflected with sufficient intensity and the infrared receiving tube is saturated. At this time the output end of the module is at a low level and the indicating diode is lit.

Sound sensor:-It is the most sensitive to sound intensity and used to detect ambient sound intensity. When the environmental sound intensity of the module fails to reach the set threshold OUT output is high level. When the environmental sound intensity exceeds the set threshold OUT

wheel and left sprocket wheel, r is the radius of track sprocket drive wheel respectively.

The summation of linear movement velocity of the automated guided vehicle robot is calculated as follow:

v vr vl r(l r ) (4)

2 2

From this equation, it is possible to say

  • If the spinning speed of each wheel is opposite and the same in magnitude, the robot is stationary and spinning. which means v 0

  • If the spinning speed of each wheel is the same and in the same direction the robot moves straight along positive x-axes

Whereas the rotational velocity is given by

v v

outputs low level. The digital output OUT can be directly connected with the single-chip microcomputer to detect the

r l

2b

(5)

high and low point level, to detect the sound of the environment.

KINEMATIC AND DYNAMIC MODEL

  1. Kinematic model

    In this paper, autonmous guided vehicle robots are like a monitored mobile platform. There are two caterpillar tracks

    After tracking the speed of an automated guided vehicle, the next behavior that can be evaluated are the local coordinate of the tracked an automated guided vehicle robot in longitudinal and lateral motions during movements. Both longitudinal and lateral motion is described as follows [7]. The center position (x, y) and orientation of the AGV are represented by

    operated by the actuators for the motion of the mobile robot, and they are placed on both sides by a mobile robot. The kinematic mode can be described as shown in Figure 1

    x v cos ( vr vl ) cos r (

    2 2

    y v sin (vr vl ) sin r (

    r l ) cos

    ) sin

    (6)

    (7)

    2 2 r l

    (vr vl ) r(r l )

    (8)

    c 2b

    2b

    r(r l )

    x cos

    y sin

    sin

    cos

    0 2

    0 0

    (9)

    0 0 1 r(r l )

  2. Dynamic model

2b

Fig. 1 Kinematic motion of a tracked mobile robot for moving to a target Using this model the speed control, the kinematic responses of the automated guided vehicle robot during traveling need to be defined as, C is the center of a mobile platform, 2b is the length between two tracks and the length of the track is l. Vl and Vr denote the linear velocities of left and right track relative to the ground, respectively. It can be calculated from the equation of the left velocity (Vl ) and the right velocity

(V ) of the tracks are written as:

Dynamic modeling of the robot is the study of motion in which forces and energies are modeled and studied. The actuator modeling is required to find the relationship between the control signal and the mechanical system input.

The motion control system of an autonomous guided vehicle can be simplified to a DC motor motion control. In modeling DC motors and in order to obtain a linear model, the hysteresis and the voltage drop across the motor brushes are neglected, the motor input voltage, vin is applied to the

r

Vr rr

Vl rl

(1)

(2)

field or armature terminals. DC motor can be modeled based on three essential electrical components: a resistor (R), an inductor (L), and a source of electromotive force (EMF), or voltage. DC motor turns electrical energy into mechanical energy and produces the torque required to move the load to

the desired output position, , or rotate with the desired output angular speed, . The torque produced a rotational

acceleration of the rotor, depending on its rotating mass or with its inertia J, and a linear viscous damping force, bm and the rotational speed.

b viscous coefficient referred to the motor shaft

kt -torque constant

kb -emf constant.

The following nominal values for the various parameters of a DC motor used:

Vin

12V , Jm 0.0551kg.m2 ,

Fig. 2 DC motor modeling From Electrical part

di d

bm 0.188N.m / rad / sec , La 0.97mHz ,

kt 0.062Nm / A, Ra 0.56, kb 0.062V / rad/ s,

T 0 (no load attached).

PID Controller

AGV motion control often uses a PID controller, taking AGV position, speed of motor and error rate of change as the controller input, and robot position, speed of motor and direction angle as the control output. In actual systems, changes in the expected values of position, speed, and direction angle, changes in actual road conditions, deviations or changes in rotational inertia, center of gravity positions, inconsistencies between the wheels and the drive, etc., make

Vin Ri L dt kb dt

(10)

global tuning of PID control parameters extremely difficult.

In recent years, PID control has been successfully applied to

Transfer function for DC motor using the Laplace transform, and rearranging gives:

Vin (s) RI (s) LsI (s) kbs (s)

(11)

(Ls R)I (s) Vin (s) kbs (s)

From equation (11) we can express I (s) as

V (s) k s (s)

mobile robots and autonomous guided vehicles. Due to the complexity and uncertainty of the AGV operating environment, it is difficult to establish an accurate model for it, and the advantage of PID control is that it does not require the establishment of an accurate mathematical model of the controlled object. Therefore, PID control is very suitable for AGV control. In response to this problem, the corresponding

I (s) b

LS R

From mechanical part

(12)

PID controller is designed in this paper, which can reduce the difficulty of PID parameter tuning and generally improve the control accuracy and robustness of the system PID control is

T J

d 2

d 2t

  • b d

dt

kt I

(13)

commonly used in feedback control

Taking Laplace transform and rearranging, gives:

t

t

JS 2 (s) bS (s) k I(s)

Substituting (13) in equation (13) and rearranging gives:

Js2 (s)

(14)

k

k

Vin (s) kbs (s) (Ls R)

t

(15)

Fig. 3 Block diagram of AGV with PID controller.

Vin (s) (Ls

Js2 (s)

Js2 (s)

R

) kbs (s)

(16)

The mathematical model of PID controller can be expressed by the following formula

kt kt

t de(t)

From equation (16) the transfer function of the input voltage, Vin (s) , to the output angle, (s) and angular speed,

u(t) PID(e) Kpe(t) KI e( )d KD

0

(19)

dt

(s) , directly follows:

(s) k

Where, e, define for each task below, is the error between the

desired value and the output value, Kp is the proportional

G (s) t

(17)

gain, K is the integrator gain, K is the derivative gain

V (s) S{(Ls R)(JS b) k k )} I D

in t b

and t is time. The control gains used in this research are

G(s) (s) kt

(18)

obtained by tweaking the various values to obtain satisfactory

Vin (s) (Ls R)(JS b) ktkb )

Where R -resistance

L -inductance of the motor J moment of inertia and

responses. If the vehicle is driven at a constant velocity, = 0 then the control input will only vary with the angular velocity, , thus:

PID(e)

It is important to know the characteristics of the autonomous guided vehicle robot. The robot platform we are doing with autonomous guided vehicle or tracked mobile robot which has two driving DC motors. We have already done the Simulate for the DC motor. First of all, we want to test the Simulink model of DC motor with the wheel for Kp=350, Ki=0 and Kd=0. The result of the simulation indicates the robots linear speed is not stable, When it is compared with the desired robot linear speed there is an error as shown in the figure belowsimulation result.

Fig 3. Simulation result for motor with Kp=300, Ki=0, and Kd=0

The linear speed time graph has high overshoot, have high undershoot, moderate rise time and no settling time and the system is not stable.

Second we want to use /test the Simulink model of DC motor with the wheel for kp=300, ki=30 and kd=0, the result of the simulation indicates that the robot linear speed is not stable but as compare the first simulate it is better as shown in the figure below.

Fig 4. Simulation result for motor with Kp=300, Ki=30, and Kd=0

It is an important thing to tune the PID parameters (Kp, Ki, and Kd) to control the control parameters and to adjust the linear speed of the robot. After many trials, with Kp=350 Ki=300, and Kd=50, the rise time and overshoot are reduced, the settling time is reached after 0.4 seconds and the system is more stable.

Fig 5. Robot linear speed time graph with Kp=350, Ki=300, and Kd=50

Experimental resultThe motion of crawler robot step experiments is conducted to verify the performance of our prototype. The maximum PWM generated by the motors controller the speed is 0.5m/s and this is the maximum possible speed of the robot.

The master controller sends instructions to each actuator or arms through servo units. Lithium battery power supply (Li- Po battery)/2200mAh, 7.4v/ is embedded to the robot for supplying electric power.

In this experiment laptop operator control unit or personal computer is used as a master controller, The unit can also store manipulator arm poses and display robot location, orientation and its battery life. It also displays keyboard shortcuts to operate and control the robot without using a hand controller. The USB cable is used for connecting the robot to pc.

As we can observe, in the figure below the AGV is moving forward when there is no obstacle in front of it. When it approached the obstacle it changes its direction as attached herewith in the figure below.

controlling AGV speeds, it can move forward, backward, right, left. We have used different sensors to track the tracking control system and avoiding obstacles. In the future work, there are many tasks to be considered the goal-seeking issue, GPS can be used to find the current coordinates and assign to desired goal position coordinates. accurate control of the heading angle and orientation can be performed using a magnetometer.

Fig.7 Motion control system of an AGV robot

CONCLUSION

In this paper, an automated guided vehicle has been controlled of the independently driven wheels is based on a kinematic model and dynamic model. An automated guided vehicle robotic platform and kinematic based on a motion tracking system are designed. The DC Motor Simulink Modeling is tested based on DC motor parameters and PID parameters, To reduce the error, the DC motor speed is controlled by the PID controller for the AGV motion control system to the desired linear speed in a robotic platform. After

REFERENCE

[1]. Zhang, Jie; Peng, Yuntao; Hung, William N. N.; Li, Xiaojuan; Tan, Jindong; Shi, Zhiping. A Case Study on Formal Analysis of an Automated Guided Vehicle System. J. Appl. Math. 2014.

[2]. Junyan Xu, Peiren Zhang, Real-time path tracking control of mobile robots based on time-varying feedback Backstepping and PID control method, Motor and Control. vol. 1, 2004.8, pp. 35-38.

[3]. ] Kanayama I, Yuta S, Vehicle Path Specification by a Sequence of Straight Lines, IEEE Journal of Robotic and Automation, vol. 3 1988.4, pp. 265-276.

[4]. Amidi O, Integer Mobile Robot Control, Pittsburgu: Carnegie Mello University Robotics Institute Technical Report CMU-RI-TR-90-17, 1990, pp. 27-31.

[5]. Kanayama I, Miyake N, Trajectory Generation for Mobile Robots, Robotic Research, 1986.3,pp. 333-340.

[6]. Lee, S. U., Gonzalez, R., & Iagnemma, K. (2016, May). Robust sampling-based motion planning for autonomous tracked vehicles in deformable high slip terrain. In Robotics and Automation (ICRA), 2016 IEEE International Conference on (pp. 2569-2574). IEEE.

[7]. Sorniotti A., Barber P., De Pinto S. (2017) Path Tracking for Automated Driving: A Tutorial on Control System Formulations and Ongoing Research. In: Watzenig D., Horn M. (eds) Automated Driving. Springer, Cham.

Leave a Reply

Your email address will not be published. Required fields are marked *