Research about sensors
Jaaay89
2535b071.pdf
Research on Robot Obstacle Avoidance Method on PSD and Sonar Sensors
Yan-rong Hou Department of Computer Science and Technology
Yanbian University Yanji, Jilin, China [email protected]
Jin-xiang Wang*(Corresponding Author) Department of Computer Science and Technology
IIP Lab. Yanbian University Yanji, Jilin, China
Abstract—In this papaer, a novel mobile robot obstacle avoidance method based on multi-sensor was proposed. According to the distant and near information of obstacles that detected by three front-end sonar sensors and four PSD sensors respectively, after analyzing the information data, obstacles distance and azimuth information were obtained, further, making strategy in the direction of motion and movement speed respectively according to the azimuth information and distance information of obstacles. So, robot could realize the autonomous patrol through the finite state transition. Finally, experiments based on AS-R mobile robot platform shows that the method is feasible and reliable.
Keywords-mobile robot; obstacle avoidance; sonar sensor; PSD sensor; finite state transition
I. INTRODUCTION A good mobile robot increasingly must have a keen
environmental perception. Robot obstacle avoidance is the essential link of robot motion, only with well avoidance work did robot motion autonomous is feasible in a more complex environment. Obstacle avoidance is the back bone of autonomous control as it makes robot able to reach to destination without collision.
In recent decades, many obstacle avoidance algorithms have been proposed, including Bug algorithm [1], Artificial Potential [2], Vector Field Histogram [3], Fuzzy Logic [4], virtual simulation platform [5] and multi-sensor information fusion technology [6-7] and so on.
Currently the application research of PSD on the mobile robot is relatively few [8], some people studied PSD as the robot's whisker. Ultrasonic sensor is one of the commonly used sensors for robot obstacle avoidance and distance measurement [9]. Han and Hahn firstly used two sets of ultrasonic sensors to detect the surface shape and position of the obstacles [10]. After that, many domestic and foreign scholars also did the research of ultrasonic sensors for obstacle identification.
II. ROBOT AND SENSORS
A. The AS-R Robot Platform The AS-R robot platform was used in the experiments.
There are two active wheels in front of it and a driven wheel
in the back of it. The wheel speed is mainly set by controlling parameters of the wheel motor, if the speed of two wheels are the same, the robot will go straight along line, otherwise, the robot will make a turn. Four PSD sensors are arranged in the front of the robot and three sonar sensors are located between the four PSD sensors. By setting the parameters of interface, the maximum working range and sampling cycle and other related information of the sonar and PSD can be controlled.
B. PSD Sensor
Figure 1. PSD sensor
Position Sensitive Detector sensor is a kind of photoelectric device which is very sensitive to incident light position.
As shown in Figure 1, the specific principle of the device is adopted reflection of light. The incident light passing through the lens L1, then, focusing on the surface of the object to be measured. The reflected light passing through lens L2, then, focusing on one-dimensional PSD sensors, finally, the light becomes into a light-spot. In the paper, the distance between two centers of lens L1 and lens L2 is defined as b, the distance between lens L2 and PSD surface is defined as f, the distance between the light-spots of the PSD and L2 is defined as x. According to the characteristic of the similar triangles, the distance between lens L1 and the object under measured can be calculated as D: xbfD /= (1)
Therefore, the value of D can be easily get by knowing the value of x. Infrared sensor has its advantage of high accuracy, the fast-speed reaction and good direction, but less
x f
B b A L2 L1
P
D
C D light
2016 3rd International Conference on Information Science and Control Engineering
978-1-5090-2534-3 /16 $31.00 © 2016 IEEE
DOI 10.1109/ICISCE.2016.231
1071
2016 3rd International Conference on Information Science and Control Engineering
978-1-5090-2534-3 /16 $31.00 © 2016 IEEE
DOI 10.1109/ICISCE.2016.231
1071
2016 3rd International Conference on Information Science and Control Engineering
978-1-5090-2535-0/16 $31.00 © 2016 IEEE
DOI 10.1109/ICISCE.2016.231
1071
than that, it is subject to environmental influence, and its detection distance is near.
C. PSD Sensor Sonar sensor is one of the main sensors for mobile robot
in indoor patrol to perception of the environment. It mainly uses the time interval between the time send out sonar sound and the time receive feedback sound which encountered the latest obstacles to calculate the distance between the sound objects and obstacles. If the time interval is set to t seconds, voice transmission distance is defined as v m/s, the distance s can be measured as s = t*v /2, so the distance between the mobile robot and the obstacle can be obtained. Rely on multi-sensors detecting different distance from the object for determination of azimuth information. The characteristics of it are high-resolution, identify transparent, irreflexive, and dark object. It also can work under atrocious conditions and easy to be integrated. The disadvantage of it is that all the ultrasonic sensors have measuring blind area, which cannot detect near obstacles. The application of sonar navigation is widely studied in reality.
Figure 2. the determination of distance and azimuth
As shown in Figure 2, A stands for the sounding point that left sonar send out, B stands for the sounding point that right sonar send out, the dotted line OX is the perpendicular bisector of the line AB, C stands for the obstacle. According to Helen theorem, the area of triangle ABC namely S ABC can be calculated. Due to the area of triangle AOC is half of the area of triangle ABC, reverse apply the Helen theorem can calculate the distance of line OC. At this time, three sides of triangle AOC are all got, according to the cosine theorem and inverse trigonometric functions, the triangle of AOC namely AOC can be obtained, which defined as
XOC=(�/2) AOC . So the azimuth and distance of point C is measured.
D. Utilize the Complementary Information Obtained from Two Sensors For the complexity and uncertainty of the environment
and the object, the information perceived by single sensor is very limited. Due to its shortage by their own hardware, sometimes the perceived information is not accurate enough. Compared with the single sensor, the multi-sensor can overcome the shortcomings such as the obtained environment information is often local and one-sided. Because of two sensors work at the same time, the noise has a big influence on the measurement value. In this circumstance, we adopted rapidly repeated sampling method to get the average value to determine the measurement data, and the maximum distance can be negligible.
Therefore, ultrasonic and infrared sensors based mobile robot measuring system were used in this paper. The sonar
sensor has strong perceive ability of the far distant objects, while the infrared sensor has strong perceive ability of short- range objects. So using combined infrared sensors and ultrasonic sensors can make up for the shortcomings of both, which would make the robot have more accurately sensing range.
III. ROBOT OBSTACLE AVOIDANCE STRATEGY
A. Data Procession Section of the Information Received by Sonar Sensor from Remote Objects Analyzing the feedback from sonar sensor to make
different control commands. If robot is far from obstacles, the moving speed of the robot will relatively quickly, the distance is closer, the speed will be slower. As shown in Figure 3, the min value is set of 80cm, the max value is set according to different environments, but not more than 7m. If the obstacle is in zone 1, the robot will go straight forward. If the obstacle is in zone 2 and zone 3, adjust the linear velocity and angular velocity of robot according to the distance and azimuth of the obstacle. If the obstacle is in zone 4, the robot must immediately reduce to slow speed and according to the data measured by the PSD to make obstacle avoidance strategy. The obstacle avoidance strategy is specific given as following :
Figure 3. sonar scanning region partition
Command 1: the obstacle is in region 1, the left, middle and right these three sonar sensors detected the distance from obstacles are all greater than the preset max value. In this case, give the command to make robot move forward quickly;
Command 2: the obstacle is in region2, the left, middle and right these three sonar sensors detected the distance from obstacles are all less than or equal to the preset max value. And the distance that the left sonar detected is smaller than the right sonar. In this case, adjust the linear velocity and angular velocity of robot according to the distance and azimuth of the obstacle. The distance is shorter, the line speed is slower, while the azimuth angle is smaller, the angular velocity will become faster. Making robot turn right by these rules;
Command 3: if the obstacle is in region 3, by the rules mentioned above to adjust the robot turn left. Additionally, if the obstacle is in the center, that is, the obstacle is in zone 2 and zone 3 at the same time,the robot should turn right;
Command 4: the obstacle is in zone 4, the left, middle and right these three sonar sensors detected the distance from obstacles are all less than or equal to the preset min value. Because that the robot is very easy collide with the obstacle,
107210721072
so we called this case as emergency state. To avoid the collide, the robot was controlled to move backwards a distance, then open PSD control mode to move forward slowly.
B. Data procession section of the information received by PSD sensor from near the distance objects Due to the control principle of PSD sensor is similar to
sonar sensor, so we introduce it brief in the paper. The one- dimensional PSD system is used in the experiment, so only a point was perceived from a PSD. PSD sensors were placed in each 30 degrees. So on a range of 90 degrees, there are in total four points were obtained. Since these four point information were important, the distance between the robot and the target can be accurately known by them. But at this case, the distance between robot and obstacle is relatively close, the speed of robot is relatively slow. Supposing the closest distance threshold is set as low. The PSD data from left to right is stored in the P1, P2, P3, P4. The control commands are as follows:
When in the PSD control mode, if the measured distances are all greater than 80cm, then stored the value of 80cm to P1, P2, P3, P4. Or else, if more than one of the measured distances are less than 80cm, then the value is controlled by the PSD sensor.
Command 5: when P1, P2, P3 and P4 are all greater than the value of low, but less than 80cm, the robot move forward slowly;
Command 6: when the leftest side P1 is less than the value of low and P2, P3 and P4 are all more than low, the robot turn right slowly;
Command 7: when P2 is less than the value of low, the robot turn right a little fast;
Command 8: when the rightest side P4 is less than the value of low and P1, P2 and P3 are all more than low, the robot turn left slowly;
Command 9: when P3 is less than the value of low, the robot turn left a little fast;
Command 10: when P2 and P3 are all less than the value of low, the robot move backwards a distance and adjust into the sonar control mode.
C. Robot obstacle avoidance method controlled by information fusion There are lots of mobile robot obstacle avoidance
strategies, the specific experimental procedure of the algorithm applied in this paper as follows:
(1) did initialization and then assigned the value of sonar to the variable now_s, and record the start time to patrol and interval time.under the state of the sonar detection. The robot received control commands from 1 to 4 according to the detected information ; when it received command 4, switch to execute procedure (2).
(2) assigned PSD to the variable now_s, during the time, the program enter the PSD control mode, The robot received control commands from 5 to 10 according to the detected information; when it received command10, switch to execute procedure (1).
(3) judged whether the interval time is up during robot patrol each time, when the time is up, stop the robot and end the robot patrol operation, end the example program created, finally release memory space and end the program.
IV. EXPERIMENTS RESULTS AND ANALYSIS Here, measurement range of the sonar is between 80cm
and 7m; the accuracy of the measurement value is 5%; frequency of measurement is set as 20HZ. Relatively larger and rough obstructions are chose in the experiments. The measurement range of PSD in actually can range from 10cm to 80cm; its accuracy is 5%; its measurement frequency is also set as 20HZ. The PSD sensors are used to measure the close objects to ensure the safe of the robot.
The experimental results show that the algorithm is feasible. The robot can independently reciprocating walking in the rectangle hallway or channel that has obstacles on both sides. The robot can avoid obstacles that on both sides and in front of it. The robot using two kinds of sensors to expand its detection range to improve the control precision, so it has no problem walking independent in certain area. In the experiment, by adjusting strategies, for example, maintain objects in a certain range of the left front, when there are no obstacles ahead, the robot go forward, when its left and front all have obstacles, the robot turn right can achieve the purpose of patrol along the wall. there are some other strategies need to be further studied. Figure 4 shown the tracing image of robot patrol in the experiment. Figure 5 shown the tracing image of the robot patrol along the wall.
Figure 4. Tracing image of robot patrol in the experiment
Figure 5. Tracing image of the robot patrol along the wall
Many experiments proved that the robot move slowly can control the state relatively easy and accurate, otherwise, move in slightly faster will cause a few errors. The robot is not much sensitive to detect the relatively small objects, such as fine wire. A great cause for the robot encounter obstacles in sometimes is some limitations exist in measuring sensor.
V. CONCLUSIONS This paper present a control strategy for mobile robot
based on PSD sensors and sonar sensors. Experimental results shows that the robot can avoid obstacles autonomously in a certain space and can patrol along the wall in the channel. Due to various reasons, the control is not
107310731073
very accurate. Though its measurement range become lager than only use sonar sensors, the perceive of object is also limited in a plane, not be enlarged to a vertical measuring range. when the object is small, occasionally miss detection an not be avoided. The experimental environment in the paper is relatively objective, so some complex environmental conditions are not in consideration.the robot can complete patrol and avoid obstacle missions in the general corridor and indoor.
REFERENCES
[1] M. Zohaib, M. Pasha, R. A. Riaz, N. Javaid, M. Ilahi, and R. D. Khan, “Control Strategies for Mobile Robot With Obstacle Avoidance,” Journal of Basic and Applied Scientific Research, 2013, 3(4), pp. 1027-1036
[2] V. Sezer, and M. Gokasan, “A novel obstacle avoidance algorithm: Follow the Gap Method,” Robotics and Autonomous Systems, 2012, 60(9), pp. 1123 1134
[3] J. Oroko, and B. Ikua, “Obstacle Avoidance and Path Planning Schemes for Autonomous Navigation of a Mobile Robot: A Review. Sustainable Research and Innovation,” Proceedings of the 2012 Mechanical Engineering Annual Conference on Sustainable Research and Innovation, 2012, 4
[4] S. H. A. Mohammad, M. A. Jeffril, and N. Sariff, “Mobile robot obstacle avoidance by using Fuzzy Logic technique,” IEEE 3rd International Conference on System Engineering and Technology (ICSET), 2013, pp. 331-335
[5] P. Wu, S. R. Xie, H. L. Liu, and J. Luo, “A novel obstacle avoidance strategy of nonholonomic mobile robot based on virtual simulation platform,” IEEE International Conference on information and Automation, 2015, pp. 185-190
[6] B. Y. Shih, C. Y. Chen, and W. C. Chou, “RETRACTED: An enhanced obstacle avoidance and path correction mechanism for an autonomous intelligent robot with multiple sensors,” Journal of Vibration and Control, 2012, 18(2), pp. 1855-1864
[7] C. M. Casey, M. Cross, D. N. Ozick, and J. L. Jones, “Obstacle following sensor scheme for a mobile robot,” US8788092 B2, 2014
[8] Y. Xue, J. T. Yang, Y. L. Dong, J. L. Shen, R. Peng, R. B. Huang, and J. Ding, “A New Approach for Autonomous Robot Obstacle Avoidance Using PSD Infrared Sensor Combined with Digital Compass,” Applied Mechanics and Materials, 2014, Vols. 511-512, pp. 101-104
[9] X. G. Ruan, and W. B. Li, “Ultrasonic sensor based two-wheeled self-balancing robot obstacle avoidance control system,” IEEE International Conference on Mechatronics and Automation (ICMA), 2014, pp. 896-900
[10] Y. J. Han, and H. Hahn, “Localization and classification of target surfaces using two pairs of ultrasonic sensors,” Robotics and Autonomous Systems, 2000, 33(1), pp. 31 41
107410741074
