CA3146577A1 - Mobile collaborative robot with obstacle avoidance - Google Patents

Abstract

This product was designed to scan and inspect the floor of a predefined space. Essentially, a camera at the end of an arm mounted on a mobile base scans the environment as the robot moves. This mobile robot is autonomous, able to avoid obstacles. The autonomy of this robot is ensured simultaneously by 4 control processes: 1. by the triangulation method using 4 routers placed at the 4 corners of the mobile surface of the robot; 2. By proximity sensors arranged circularly around the robot; 3. By scanning the surface where the movement is performed 4. By precise positioning encoder sensors placed on the drive motors. Because the four types of sensors work simultaneously in real time, the positioning is accurate.

Description

Mobile collaborative robot with obstacle avoidance INTRODUCTION
In today's world of robotics, fully autonomous robots have not yet been created, because even advanced robots have not yet been upgraded for a complete autonomy.
As a potential solution, this mobile robot system is presented. This system stands out from other high-end robots with its revolutionary characteristics.
Unlike other robots, which use 2-3 sensors, this system uses 8 ultrasonic sensors and a LiDAR module, which all provide 360 coverage and high obstacle avoidance precision (Fig. 1). Moreover, a camera is equipped with the system, serving to scan the environment, to send a video stream direct to a web server and capture individual frames (images) when invited by a program that meets the needs of the user. All entries are sent to a web server at from which a complexly designed AI algorithm processes the data and command with steering precision. Despite receiving data from several sources, the AI processes the instructions instantly. With 100% autonomy, this system opens several practical applications and it can also collaborate with other robots in order to diversify its features through a network of devices oriented towards a common goal. For example, the tank of jet fuel from a aircraft is normally examined by human beings, and they are required to work in spaces restricted and to inspect the full range of the tank, which may represent hard work (Fig.2).
robot presents itself as a solution adapted to the inspection of the tanks of fuel, because it can accurately detect any faults in the tank rivets or check any traces of corrosion. In addition, it can be combined with another robot in order to be able to to be transported in different sections of the reservoir.
'Ll'44' 44-FIG. 1. Final product with erect arm Date Received/Date Received 2022-01-24 READ*
. .
I-\
-\\
-rs . =
FIG. 2. Aircraft fuel tank inspection performed by crew members TABLE I: BILL OF MATERIALS
No. Part Name Quantity ROBOT FOUNDER HAT

OF SUN

2 Founder of Sun PCA9865 1 Sun Engine Driver Module

3 1 Founder

4 SF-5R02 8 Sun Founder SFOO6CServo 1 6 DC gear motor 2 7 Raspberry Pi 4 1 8 myCobot-Pi 1 9 Aluminum bracket SF-5R02 1 myCobot Aluminum Stand -Pi 11 USB Camera 1 TF laser distance sensor Mini-LiDAR (ToF) He. MOBILITY
The product is built using a robotic 4-wheel machine, controlled using a Raspberry Pi 4 running Python scripts, which execute instructions data-driven directional received from the eight ultrasonic sensors surrounding the robot (Fig. 3 and Fig. 4), as well as from a module LiDAR (Fig. 19) for a complete view.
Date Received/Date Received 2022-01-24 õµ re ("4, , ^15tt 40 ae, ,..' . "1.; \ ,Ii'et4 e 4-4)11.:÷ 1 '' 4à t' Ye.u.Le 7:1,",, õbile ' ~JI,, efr D.
FIG. 3. Mobile base equipped with ultrasonic sensors, side view The robotic mobile is composed of three modules, as well as a servomotor and two DC geared motors. The modules are all Sun modules Founder, namely Robot HATS, PCA9865 and the Motor Driver module. The servo used is Sun's SFOO6C
Founder, and both DC motors: F130SA-11200-38V. The servo is used for control the direction of the robot by means of the axle of the front wheels, while the DC motors control the direction of the robot via the rear wheels. The HATS robot is directly connected to the Raspberry Pi via general-purpose input-output pins (GP10). By receiving the data processed via Python scripts, it relays the data to the pulse width modulation driver (PCA9865), which in turn transmits the data directly to the servo. In addition, the PWM sends data to the module order of motor which, in turn, controls the speed and direction of the motors to direct current.
Date Received/Date Received 2022-01-24 , -, -f, I
FIGURE 4 Mobile base equipped with ultrasonic sensors, top view Ultrasonic sensors coupled to the Raspberry Pi send data to a algorithm artificially intelligent via Python scripts, which in turn call the IP address of a web server based on Flask linked to AI. Once the algorithm has processed the data, it returns them to the Raspberry Pi under form of directional instructions for motors.
III. ULTONIC SENSORS
The ultrasonic sensors used are model SF-5R02 from Sun Founder (Fig.

5), and eight of them are employed. The purpose of ultrasonic sensors is to calculate distances nearby objects in order to avoid them accordingly. The sensors are connected to the Raspberry Pi via the GPIO ports the computer. These sensors are unique, compared to ultrasonic sensors conventional, because there are normally two separate physical connections to be established: one for send waves ultrasound (triggering), and the other to receive them after they bounce off an object (echo). The SF-5R02 sensors, however, implement this function in a single connection, thus simplifying the wiring.
TABLE II SPECIFICATIONS OF SF-SR02 SENSOR[2]
Operating voltage Working current 16mA

Working frequency 40Hz Maximum range 700 cm, stable 500 cm Minimum range 2 cm Trigger input signal TTL pulse 10p.S
Echo Input Signal Signal Dimensions 46x20.5x15mm lifting TTL input Date Received/Date Received 2022-01-24 Operating voltage Working current 16mA

Working frequency 40Hz Maximum range 700 cm, stable 500 cm Minimum range 2 cm Trigger input signal TTL pulse 101.15 Echo Input Signal Signal Dimensions 46x20.5x15mm lifting TTL input Ultrasonic sensors are widely used in the industries of automation as well as in medical offices. They are commonly used as proximity sensors or level sensors.
Proximity sensors are often used in the automotive industry for self-driving systems parking and anti-collision, as well as in robotics for systems detection/avoidance obstacles. Level sensors detect, monitor and regulate liquid levels.
In general, an ultrasonic sensor contains two main components: a transmitter and a receiver. A transmitter emits ultrasound using piezo crystals and a receiver meets sound after it has traveled to and from the target object.
shot ;=,,,,, //,',itta,',,e_ ,, ,, .. ._..."e '=',' he --,----,.,-- , -__ million 1111 , .-...õ,..-II
romei,/
, .7....._ ..

FIGURE 5 SF-SR02 Ultrasonic Sensor [2]
An ultrasonic sensor works by transmitting an ultrasonic pulse to 40 kHz, which is higher than the frequency of sound audible to humans (20 Hz to 20 kHz). This impulse moves through the air and bounces back to the sensor if there is an obstacle or object within its range functioning. The time taken by the pulse to be emitted and received is recorded, and the distance from the object relative to the sensor is then found through time registered with reference to the speed of sound in air. This type of measurement is commonly called measurement of the flight time (Fig.6).
Date Received/Date Received 2022-01-24 EIVIITTED SOUND WAVES
TO VOMIT -DETECIOR = õfr-', FIG. 6. Travel time measurement IV. PYTHON SCRIPTS
This system uses 3 python scripts that run on the Raspberry Pi. In order to obtain the distances read by the 8 sensors (Fig.7), 2 python programs were created. The first program reads data sensors 1 to 4 and the second program reads data from sensors 5 to 8, and the two programs run recursively with virtually no delay.
Sensor 4 Sensor 3 Sensor 5 From Sensor 2 Sensor 6 Rea Sensor I Sensor 7 Sensor 8 FIG. 7. Eight-sensor configuration for the AI algorithm Two programs were used instead of one program for the 8 sensors so that the data are read faster from the sensors. Indeed, with a program, each sensor must wait for 7 more sensors before you can read values again, then that with two programs, each sensor only waits for 3 other sensors. These programs are almost identical to each other others, and they each serve to read the value of their respective sensors before sending the data to their own text files, as string values. Each file text contains data Date Received/Date Received 2022-01-24 read by 4 sensors (the first text file is for sensors 1-4, and the second, 5-8), and the files text are overwritten with each iteration of the program.
Al server Flask webserver raspberry-pi Robot HAT
14 \
Ultrasonic PWM
sensors ________________________ Engine Servo drives driver ______________________________ (forehead) DC motors (back) Figure 8: Communication flowchart of mobile robot and AI server The programs use the "Ultrasonic Avoidance" module, which is a module default provided by SunFounder which calculates the distance of objects read by the sensor. this is accomplished by calculating the time necessary to receive a pulse signal once emitted.
The distance is first read by the sensor and then assigned to a variable.
Then, in order to avoid errors logical, the distances read by the sensors are only considered if they are positive. Once the positive values taken, they are then checked if they are within in the range from 0 to 80 centimeters.
Any value greater than 80 centimeters is reduced to 80 centimeters, which is the upper limit defined in the system. If the value is within the defined range, it is converted to a value of string with a space ( ) and a comma ( , ) appended. If she was converted to 80, the variable takes 80 as value and is then converted to a string value with space and the comma suffixed.
Once all four sensors have read a value and have a string that represents the read value, they are then concatenated and sent to the text file.
Date Received/Date Received 2022-01-24 While the sensors read the data and send it to files text as values chain, a third script will also run concurrently and serve means of communication between mobile robot scripts and web server. This program first reads both text files and stores each value as a string. The chain will be at format "12,34,56,78,90,12,34,56". It is then concatenated to a string containing the URL of the web server. There final string after concatenation will look like: "IP_address: 9999 / predict /? Sonar =
12,34,56,78,90,12,34,56". The value returned by this URL determines the direction in which the robot will move in response to readings from its environment (see flowchart shown in Fig.8). THE
program executes the aforementioned commands recursively.
V. WEB SERVER AND AI
The web server is created with a Python script that uses the Flask framework.
The server script receives the data that the Python scripts of the Raspberry Pi send to it by calling its IP address. The server web is coded to then relay the data to an AI server, which processes the data to calculate directional controls. Orders are then sent back via the raspberry web server Pi (sent to its IP address), where they are received by a Python script that indicates the exit direction of the motors (refer to the flowchart in Fig.8). THE
directional data returned to the module are in the form of strings, and the three possible values are "0", "1"
and "2". The "0" tells the module to move forward, while "1" and "2" tell the move module left and right respectively.
Explicit details of the math and code behind the AI algorithm can not be fully disclosed due to copyright issues.
Essentially the web server receives the data sent by the eight ultrasonic sensors, the LiDAR and the camera mounted on the cobot.
It then converts said data into string values and then sends it to the AI server. Al server processes the data and sends it back to the server as commands directional.
VI. LIDAR
Once the mobile robot equipped with an ultrasonic sensor is able to avoid the obstacles of autonomously by communicating with the web server and the AI, a sensor of detection and Light ranging (LiDAR) is added to the model. The sensors at ultrasounds are useful for detecting low-level objects closer to the ground, but do not detect at a height enough for a optimum object avoidance. A LiDAR mounted on top of the mobile robot responds to this requirement (as seen in Fig. 9 to the right of the camera).
LiDAR sensors are used to provide remote sensing and more accurate object distances while being able to visually map an environment using of a software image detection.
In addition to ultrasonic sensors, LiDAR sensors are another source information for the AI algorithm that would make obstacle detection more accurate.
Date Received/Date Received 2022-01-24 e -naked 4ieetILM.;"
4.µ
,---FIG. 9. Full robot with TF Min' LiDAR mounted top right support slab LiDAR sensors are similar to ultrasonic sensors, except that the LiDAR sensors use pulses from a laser to collect measurements.
LiDar sensors also have a transmitter and a receiver. The transmitter normally send several pulses per second and the receiver reads the light waves that bounce off an object.
LiDAR sensors, like ultrasonic sensors, also use flight time measurements, except that the speed of sound in air is replaced by the speed of light in the air, since lasers are used (Fig.11). The system currently uses the laser sensor unidirectional TF Mini LiDAR
(Fig.10) as a prototype to test the functionality of Al algorithm.
It looks like the sensors to ultrasound used in the system, but it is much smaller.
4 , ree, =='`),1 : e' r e&V-e=====e to there = ' ==
=r=
) , e FIGURE 10 TF Man LiDAR (ToF) Laser Distance Sensor Date Received/Date Received 2022-01-24 TABLE III: TF MINI LiDAR SPECIFICATIONS
Working voltage 4.5-6V
Average power 0.6W
UART(TTL) Communication Interface Frequency of 100Hz refreshment Maximum range 12m Minimum range 0.3m Field of view angle 2.3 Accuracy ranging 1%(<6m), 2%(6-12m) Size 42x15x16mm e, 0 0.3111 e. 3m 7m 12m g FIG. 11. TF Mini LiDAR Series Diagram VII. MYCO BOT-PI
The myCobot-Pi is a 6-axis collaborative robot (Fig.12). He adopts a Raspberry Pi 4 microprocessor with Raspbian by integrating it into its base, and it is used to program the movement of the cobot. In order to control the movement of the cobot, Python code is used. The robot allows also the programming via the Robotic Operating System (ROS) which is built into its microprocessor RaspberryPi.
The cobot serves as the arm of the mobile robot system and is mounted on the base mobile. For the needs of this product, the cobot is programmed to expand horizontally with its end effector directed towards the bottom. It rotates clockwise and anti-clockwise.
counter clockwise shows between two user-defined positions, similar to a pendulum swinging towards a lateral gravitational attraction. A camera is mounted at the end of the cobot, which allows a environment analysis.
Date Received/Date Received 2022-01-24 Fig. 12. myCobot-Pi VIII. THE CAMERA
The camera used to scan the environment is a Raspberry USB webcam Pi (on the side left of Fig. 13). The camera cable is plugged into one of the ports USB from the Raspberry Pi to the base of the myCobot-Pi. The lens itself is mounted above the effector cobot terminal, so so that it points in the same direction as a protruding clamp outward from the end effector.
The camera streams visual data to a configured local web server through Motion software, running on the cobot's Raspberry Pi. The web server can be rendered accessible to the public through the port forwarding, the configuration of which depends on the configuration of the user's router and security needs.
11111111111.151PINMIMO e,19 =
ddi ko=;
01111111lb Fig. 13. Camera (at u.uche) and mmt-test LIDAR (at right) attached eeffeMeur myCobot-Pi terminal The camera can be programmed to stream data for that the script of robot movement runs. It can also be programmed to take a picture during a Date Received/Date Received 2022-01-24 predefined trigger event (such as a sudden color change) and to store the images in a local directory within the cobot's Raspbian system, or even send it to a server through the Raspberry Pi.
IX. POSITIONING BY TRIANGULATION
The mobile robot has at least 2 Raspberry Pi microprocessors, each with own address IP. In order to accurately locate the position of the robot at all times, the Wi-Fi positioning can to be used.
This geolocation system works by reading the signal strength of the mobile robot via appropriately placed routers.
The first step in triangulation is to find the position horizontal of the mobile robot. This is made by considering two points and forming a triangle with the robot mobile as third point. This will be used to find the position of the mobile robot between two routers.
By comparing the strength of mobile robot signal on both routers, we can find the proximity of the mobile robot by relative to each router. The distance can then be approximated from signal strength. By example, in Fig. 14, the mobile robot is closer to Router 1 than to router 2 as shown by the opacity of the colors representing the strength of the connection.
The problem with using only two routers is that the mobile robot can be on either side of the line that connects the Router 1 and Router 2 (as shown in Fig.14). Therefore, a another router is used and the signal strength of the mobile robot with the router 3 indicates the position by relative to the common line between Router 1 and Router 2.
=
End point of Router 1 work space (xi y1 z1) , , g '-Mobile where Robot (x0,y0,z0) =
=
Router 2 (x2,y2,z2) di>Origin(0,0,0) Fig. 14. Triangulation with 2 wifi routers. The color of the connection represents signal strength (green-good, orange-ok, red-weak) Date Received/Date Received 2022-01-24 The second step is to find the height of the robot and it is calculated of the same way than the horizontal distance, using the three routers and their approximation of the strength of the distance signal. A fourth router Router 4 (Fig.15) is used to a closer approximation accurate horizontal and vertical distance, as with three routers, the approximation of the location is only 4 to 5 feet (1.2 to 1.5 m).
Fig. 15. Triangulation with four wifi routers. The color of the connections represents the power of signal (green: good, orange: ok, red: weak) In order to accurately calculate the distance between the robot and the router, the signal strength in dBm or RSSI
of the router is used. First, a fixed value of the intensity of the signal is assumed which will be used to interpolate the distance. Then the robot is moved so that routers provide the same reading. Finally, the distance from the robot to the specified signal strength of each router is measured. For example, consider a fixed intensity of -50 dBm. The robot is then moved to a position where Router 1 reads - 50 dBm. The distance between the robot and the router 1 is now measured and noted. The same process is repeated with Router 2, Router 3, and Router 3.
router 4.
X. SCANNING THE ENVIRONMENT USING A 2D LIDAR SENSOR
Another reason for the use of LiDAR sensors (Fig.16) is their ability to emit impulses at 3600, which means it can scan the entire environment surrounding the mobile robot, including potential blind spots that ultrasonic sensors may not detect ( Fig.19).
Before being entrusted with performative tasks, the LiDAR must first familiar with his environment. It does this by mapping the environment. The moving robot is programmed to follow a path approximately following the perimeter of the workspace targeted, as well as through the center of the workpiece, the specific trajectory of which depends on the size and shape of space of work. This allows the LiDAR to be exposed to the whole environment and to record all obstacles it detects, including walls and room boundaries (Fig.17 and Fig.18).
Date Received/Date Received 2022-01-24 7' =
µ.410 =
. , fl sorting Fig. 16.YDLiDAR X4 Once the map is registered, the LiDAR device can be assigned target trajectory points, whether communicates to the engines through the AI server. The AI algorithm collaborates with the LiDAR device to find the optimal path to the destination from its current location. THE
along the way, the LiDAR analyzes its environment and sends data to the server in real time. When he encounters an object obstructing its path, the AI server calculates the most efficient path to avoid the object while still directing to the target, then returns directional controls.
I iii ) I;
ïiï = , ' 1 Fig. 17. Example of a map of the workspace recorded by a module LiDAR, approx. 40m2 Date Received/Date Received 2022-01-24 IO.
3.15, front = 1.
...
THE .
/ = . .
. =
...
14 4.b.. 8:1 If , 1.1P
Fig. 18. Example of workspace map recorded by the module LiDAR, approx. 40m2 r 1 #4,e4,1 rti,,-,, ' I
I
I;
:
Fig. 19. Full view of the robot with a 360 LiDAR facing forward TABLE IV: YDLiDAR X4 SPECIFICATIONS
Working voltage 4.8-5.2V
Working current 330-380mA
Baud rate 128000 bps Frequency range 5000 Hz Date Received/Date Received 2022-01-24 Range Resolution <0.5mm(for Range<2m) <1% of distance(Range>2m) Frequency of

6-12Hz scanning PWM frequency 10kHz Scan angle 0-360 Maximum range 10m Minimum range 0.12m Angular resolution 0.48 -0.52 Accuracy ranging 1%(<6m), 2%(6-12m) Wave length 775-795nm laser 3mW laser power Size 102x71x63mm XI. FUTURE WORK
The robot is a very unique and innovative design, but still has a lot room for improvement. For example, the camera currently used in the system is very basic and does not produce high quality images or live streams that could be essential to the AI server.
Therefore, the use of better cameras that capture images of good quality and live video stream is required.
The mechanical structure currently in use is not fully capable of support myCobot Pi at the top, and the movement of the mobile robot is not smooth. Moreover, some parts of the structure use plastic, which is not durable. Therefore, a upgrade should be brought to the structure which will use aluminum or steel in order to be able to bear loads useful while not impeding the movement of the robot.
Although the sensors currently used for the system are efficient, they are not durable and have a smaller field of vision, which may increase sensitivity to blind spots during movement. Therefore, bigger and better sensors will be used to send data to the AI server more efficiently and accurately.
For the mobile robot to inspect the fuel tanks, it requires the help of another robot able to scale walls, as jet fuel tanks are partitioned and the robot must move effortlessly through the openings in each partition.
The time it takes for the mobile robot to communicate with the web and AI server is minimal for moment. However, it can be much lower. The use of best sensors and motors capable of sending and receiving commands much faster than those currently used, and better internet connectivity are ways to minimize the latency.
Date Received/Date Received 2022-01-24 The AI model that calculates direction based on sensor input is still in progress development and optimization to ensure that data processing takes less time and that steering commands are more precise and instantaneous.
XII. EQUATIONS
The mathematical formula used by ultrasonic sensors to calculate the distance from a target is the next one :
The distance is the product of the duration of the pulse, which is the time taken by the impulse to be transmitted and received, and the speed of sound. This equation is multiplied by 0.5 because the duration of the pulse provides total flight time, while distance only provides length of an object from the sensor.
The formula used to calculate the distance from a target object to a sensor LiDAR is as follows:
The distance in this equation is found in the same way as in the previous equation, but in this case, the distance of an object from the LiDAR sensor is measured using the speed of the light.
Date Received/Date Received 2022-01-24

Claims

Mobile collaborative robot with obstacle avoidance Claims 1. The mechanical design of the mobile robot, using a structure original mechanics, using toothed belt transmissions and motors with integrated driver, with forward-backward movement independent of the two motors.
2. Method of using 4 kinds of control sensors simultaneously and in real time, with correction real-time position:
2.1 encoding positioning Optical incremental encoders are located on the drive motors electrical. They measure the movement in a very precise way and move the robot with more precision, after it has been positioned less precisely by the other methods: triangulation, Lidar and proximity detectors.
2.2 positioning by proximity detectors This system uses 3 python scripts that run on the Raspberry Pi. In order to get the distances read by the 8 sensors (see details in the file description or in point 3 of this document).
2.3 positioning by triangulation The mobile robot has at least 2 Raspberry Pi microprocessors, each with own address IP. In order to accurately locate the position of the robot at all times, the Wi-Fi positioning can to be used.
This geo-location system works by reading the signal strength of the mobile robot via appropriately placed routers.
The first step in triangulation is to find the position horizontal of the mobile robot. This is made by considering two points and forming a triangle with the robot mobile as third point. This will be used to find the position of the mobile robot between two routers.
By comparing the strength of mobile robot signal on both routers, we can find the proximity of the mobile robot by relative to each router. The distance can then be approximated from signal strength. By example, in Fig. 14, the mobile robot is closer to Router 1 than to router 2 as shown by the opacity of the colors representing the strength of the connection.
The problem with using only two routers is that the mobile robot can be on either side of the line that connects the Router 1 and Router 2 (as shown in Fig.14). Therefore, a another router is used and the signal strength of the mobile robot with the router 3 indicates the position by relative to the common line between Router 1 and Router 2.

End point of Router 1 work space (xl ,y1 ,z1 ) , 111>
, rrror rvmobile Robot e (xo,yo,zo) f Router 2 (x2,y2,z2) =Origin(0,0,0) Fig. 14. Triangulation with 2 wifi routers. The color of the connection represents signal strength (green-good, orange-ok, red-weak) The second step is to find the height of the robot and it is calculated of the same way than the horizontal distance, using the three routers and their approximation of the strength of the distance signal. A fourth router Router 4 (Fig.15) is used to a closer approximation accurate horizontal and vertical distance, as with three routers, the approximation of the location is only 4 to 5 feet (1.2 to 1.5 m).
Fig. 15. Triangulation with four wifi routers. The color of the connections represents the power of signal (green: good, orange: ok, red: weak) In order to accurately calculate the distance between the robot and the router, the signal strength in dBm or RSSI
of the router is used. First, a fixed value of the intensity of the signal is assumed which will be used to interpolate the distance. Then the robot is moved so that routers provide the same reading. Finally, the distance from the robot to the specified signal strength of each router is measured. For example, consider a fixed intensity of -50 dBm. The robot is then moved to a position where Router 1 reads - 50 dBm. The distance between the robot and the router 1 is now measured and noted. The same process is repeated with Router 2, Router 3, and Router 3.
router 4.
2.4 Positioning by scanning the environment using a lidar sensor Another reason for the use of LiDAR sensors (Fig.16) is their ability to emit impulses at 360, which means it can scan the entire environment surrounding the mobile robot, including potential blind spots that ultrasonic sensors may not detect ( Fig.19).

Before being entrusted with performative tasks, the LiDAR must first familiar with his environment. It does this by mapping the environment. The moving robot is programmed to follow a path approximately following the perimeter of the workspace targeted, as well as through the center of the workpiece, the specific trajectory of which depends on the size and shape of space of work. This allows the LiDAR to be exposed to the whole environment and to record all obstacles it detects, including walls and room boundaries (Fig.17 and Fig.18).
Once the map is registered, the LiDAR device can be assigned target trajectory points, whether communicates to the engines through the AI server. The Al algorithm collaborates with the LiDAR device to find the optimal path to the destination from its location current. Along the path, the LiDAR analyzes its environment and sends data to the server by real time. When he encounters an object obstructing its path, the Al server calculates the path more effective in preventing the object while moving towards the target, then returns commands directional.
3. The way of detecting and circumventing obstacles, by the mobile robot, with a model that uses 8 proximity sensors This system uses 3 python scripts that run on the Raspberry Pi. In order to obtain the distances read by the 8 sensors (Fig.7), 2 python programs were created. The first program reads data sensors 1 to 4 and the second program reads data from sensors 5 to 8, and the two programs run recursively with virtually no delay.
Two programs were used instead of one program for the 8 sensors so that the data are read faster from the sensors. Indeed, with a program, each sensor must wait for 7 more sensors before you can read values again, then that with two programs, each sensor only waits for 3 other sensors. These programs are almost identical to each other others, and they each serve to read the value of their respective sensors before sending the data to their own text files, as string values. Each file text contains data read by 4 sensors (the first text file is for sensors 1-4, and the second, 5-8), and the files text are overwritten with each iteration of the program.

The programs use the "Ultrasonic Avoidance" module, which is a module default provided by SunFounder which calculates the distance of objects read by the sensor. this is accomplished by calculating the time necessary to receive a pulse signal once emitted.
The distance is first read by the sensor and then assigned to a variable.
Then, in order to avoid errors logical, the distances read by the sensors are only considered if they are positive. Once the positive values taken, they are then checked if they are within in the range from 0 to 80 centimeters.
Any value greater than 80 centimeters is reduced to 80 centimeters, which is the upper limit defined in the system. If the value is within the defined range, it is converted to a value of string with a space ( ) and a comma ( , ) appended. If she was converted to 80, the variable takes 80 as value and is then converted to a string value with space and the comma suffixed.
Once all four sensors have read a value and have a string that represents the read value, they are then concatenated and sent to the text file.
While the sensors read the data and send it to files text as values chain, a third script will also run concurrently and serve means of communication between mobile robot scripts and web server. This program first reads both text files and stores each value as a string. The chain will be at format "12,34,56,78,90,12,34,56". It is then concatenated to a string containing the URL of the web server. There final string after concatenation will look like: "IP_address: 9999 / predict /? Sonar =
12,34,56,78,90,12,34,56". The value returned by this URL determines the direction in which the robot will move in response to readings from its environment (see flowchart shown in Fig.8). THE
program executes the aforementioned commands recursively.
5. the artificial intelligence model to circumvent obstacles based on 8 proximity sensors arranged symmetrically around the mechanical structure. Their role is to detect obstacles real time and to modify, in real time, the trajectory of the robot, so that it do not hit objects around him as he moves. Several models have been tested before, but this one gave the best results.