AUTONOMOUS GROUND ROBOTIC VEHICLE

FIELD OF INVENTION

[0001] The present invention relates to one autonomous (unmanned) robotic vehicle configured for balancing and shock absorbing technique in rough terrains particularly agricultural fields. More particularly, the present invention discloses a high load capacity robotic vehicle with modified robust wheel balancing mechanism guided by sensors including camera and GPS sensor for path navigation.

BACKGROUND

[0002] Nowadays, automation is playing a dominant contribution in the agricultural domain. Different types of autonomous robotic vehicles are taking the place of humans in agriculture monitoring and operations. In the agriculture field, the Global Positioning System (GPS) sensor is a very effective technology for location determination. An autonomous robotic vehicle with a pesticide sprayer can move towards certain locations with the help of the GPS sensor and use a camera for obstacle avoidance and vegetable growth inspection. One problem for these types of pesticide sprayer autonomous vehicles is that they can’t identify insects in vegetables like – cabbage, cauliflower, broccoli, kohlrabi, spinach, etc., because of their camera position. Most crop monitoring vehicle has a heavy size and long height that causes some balancing problems. These types of vehicles can’t move between farming vegetables; it can cause some damage to the vegetables. Also, the power consumption is high for these types of vehicles, causing a larger battery size and vehicle size. For this purpose, an autonomous agricultural ground vehicle is designed and developed for scouting and spraying operations in a crop field. Some other examples of agricultural ground vehicles are considered hereafter with a brief explanation.

[0003] An intelligent automated vehicle with a smart visual sensor for crop management is designed in US patent no. 10,721,859 B2. The visual sensing system helps the proposed vehicle in path navigation with other operations like spraying, weeding, and seeding in a crop field. Several tasks are performed after the real-time images have been analyzed.
[0004] An automatic treatment process for weeds with the help of a robot has been shown in US patent no. 10,681,905 B2. The captured images of weed were selected and an automatic treatment tool with the said robot performs the treatment for the said weeds.

[0005] One autonomous tractor for agriculture is used in US patent no. 2019/0133024 A1. The method of autonomous agriculture in a vast agricultural field has been discussed with the help of the autonomous tractor that operates at a vast distance from the minimum supports.

[0006] A robotic vehicle with a spraying system has been discussed in US patent no. 9, 877, 470 B2. An autonomous robotic spraying vehicle with a forward-looking LiDAR sensor and self-locating GPS sensor has identified the path in agricultural land by getting directions from a wireless mobile control center. Additionally, a map per vehicle is also used for the path information in the said area.

[0007] One ground vehicle for scouting operations in a crop field is used in US patent no. 2016/0157414 A1. The ground scouting vehicle is implemented with another air scouting drone in a crop field where the said scouting devices communicate with each other and share their findings.

[0008] An agricultural robot performs multiple operations like harvesting, weeding pruning, and culling using a visual sensing system in US patent no. 7,854,108 B2. In an agricultural field, the robot also performs measuring and managing crops. The cameras in the robot system are used to identify the objects and map the plant locations.

OBJECT OF THE INVENTION:

[0009] The object of the invention is to produce an autonomously movable robotic vehicle which has an improved effectiveness.

[0010] Another object of the present invention is to autonomously movable robotic vehicle for agricultural applications, comprises of a motor to drive the autonomously movable agricultural vehicle and a battery to supply energy to the motor, and wherein the autonomously movable agricultural vehicle comprises a control circuit.

[0011] Yet another object of the present invention is to wherein the method comprises the step of adjustment of a direction of motion and speed of the autonomously movable robotic vehicle such that the energy drawn from the battery over a predefined distance is minimal.

SUMMARY OF THE INVENTION:

[0012] Therefore such as herein described there is provided a specific selection for an Autonomous Ground Robotic Vehicle comprising of a GPS sensor to navigate through the planned locations; a geomagnetic direction sensor for detecting a direction of a vehicle body;
a GPS receiver for recognizing real time travelling position of the vehicle body; an infra-red camera and ultrasound sensor configured to detect and record real-time terrain obstacles; a wheel architecture system for perfect balance on odd and even surfaces; .and a programmed processor connected with infrared camera module and coupled with L80 GPS module with an embedded patch antenna enabling said vehicle and causing vehicular wheel centre position and axle to follow a target terrain of said vehicle body.

[0013] In an embodiment there is provided an autonomous Ground Robotic Vehicle (AGRv) is developed with higher accuracy in self-position estimation on the rough terrain of the agriculture field. The robot uses plurality of sensor e.g. camera ultrasound, GPS for self-position estimation and to object / obstacle detection. The robot has its independent processor module with storage medium, GPS module. In the agriculture terrain, the geometric locations are provided by the GPS sensor to the robot.

[0014] In another embodiment, the ground vehicle is operational considering agricultural terrain having a Battery power supply operates in an agricultural field for 72 minutes on average. The ground vehicle consists of a Cortex-A72 processor coupled with storage medium that includes other sensory systems, motor drivers, motors, along with pumps, and spraying systems as optional. The wheel system of the ARGv is designed with the utmost balance that can support double the vehicle's body weight. The two legs of the ground vehicle have an angle of almost 45-degree with the wheels. This angle helps the wheels to diversify the applied force into equal amounts. Also, the springs are attached to the two legs of the ground vehicle helps in shock absorption. The maximum speed measured for the ground vehicle is 8.64km/hr and the total time measured to follow a path of 100meter with a full load is 11 minutes 7 seconds.

[0015] The autonomously movable robotic vehicle according to the invention comprises a motor to drive the autonomously movable agricultural vehicle, and comprises a battery to supply energy to the motor. The autonomously movable agricultural vehicle comprises a control circuit including Cortex-A72 processor is configured to control the motor to adjust the direction and speed of the said robotic vehicle. The control circuit is configured to set the speed of the autonomously movable robotic vehicle such that the energy drawn from the battery over a predefined distance is minimal.

[0016] In another embodiment of the autonomously movable robotic vehicle, comprises an optical sensor to recognize obstacles. An optical sensor is a high resolution camera which is mounted on the said vehicle and with which the obstacles can be recognized. The advantage of optical sensors is that they are relatively inexpensive and that image processing can possibly be carried out on the recorded images in order to collect additional information relating to an obstacle. The control circuit is programmed to read and analyze the data collected from the sensor in real time. The control circuit is, configured to add obstacles in the route to the storage medium if new obstacles are identified by the said optical sensor.

[0017] In another embodiment, the terrain features are recorded and evaluated by the said camera and compared with the threshold limits by the controller circuit. The wheels and the axle are controlled thereupon by the control circuit as per the pre-stored program.

[0018] In another embodiment, the battery power consumption of the autonomously movable robotic vehicle is related to the speed of the autonomously movable agricultural vehicle.

[0019] In another embodiment of the autonomously movable robotic vehicle, the control circuit is configured to adjust the differing speed value and/or the energy value associated with the differing speed value of the defined route in the storage medium if the energy consumption during the differing speed value is lower than the energy value stored in the storage medium enabling, a machine learning and self-adjusting vehicle.

[0020] In another embodiment of the autonomously movable robotic vehicle, the vehicle comprises a speed sensor to measure the speed of the autonomously movable robotic vehicle. The control circuit receives a real time speed signal from the said speed sensor which indicates the measured speed, wherein the control circuit uses the speed signal to determine a position of the autonomously movable agricultural vehicle along the route. The autonomously movable robotic vehicle comprises speed sensor which roughly measures the real time speed of the vehicle, from where, a location of the autonomously movable vehicle in the route can be determined precisely.

[0021] In another embodiment of the autonomously movable robotic vehicle, the said vehicle comprises a GPS sensor to determine a location point along the route. The location point of this type could preferably comprise a start point and/or end point of a route. Between the location points, the said vehicle could determine the position within the route with the aid of its speed sensor. The GPS sensor could also comprise a sensor to identify a distant obstacle and determine the distance to said obstacle, so that the said vehicle does not hit into the obstacle but, change direction or stop before it. The GPS sensor could, for example, also comprise a gyroscope.

[0022] In another embodiment of the autonomously movable robotic vehicle, the said vehicle comprises a real time energy sensor to measure the energy drawn from the battery. The said energy sensor could, for example, measure a current which is drawn from the battery. Given that the voltage over a battery remains more or less constant, the current drawn provides a measure for the energy drawn from the battery.

[0023] The working algorithm is designed with multiple parameters like – number of iterations, different positions, minimum distances, minimum angles, threshold values, etc. The RGB images are extracted from the live images captured by the camera, and then the threshold value of pixels is identified. One binary operation for masking is applied to the RGB images, and then the masked RGB images are converted into black and white images with a lower threshold value. The pixels of black and white images are counted. Then a higher pixel value is assigned for the white pixel values.

[0024] When the white pixel values are matched with the specific higher threshold value, then the desired objects are identified by the algorithm. If the insects are detected, then the algorithm instructs to perform spraying operations with the help of the pumps. If the obstacle is detected then the algorithm instructs to perform wheel rotation with the help of motor drivers. After reaching the final goal the algorithm stops iterating, and the ground vehicle stops.

[0025] One 40 cm heightened vehicle can identify insects in small size crops (cabbage, cauliflower, tomato, eggplant, etc.). The camera on the vehicle can be mounted at a 45-degree angle vertically and attached with a dc motor for 180-degree movement horizontally. The ground vehicle is used in a crop field with a path planning strategy to monitor all crops from the start position to the goal position. One GPS module is used to locate the desired positions for the monitoring facility. The ground vehicle inspects the crops and other objects with the help of the camera.

BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS:

[0026] Fig. 1: Shows the path planning for agricultural crop monitoring;

[0027] Fig. 2: Shows the autonomous algorithm for the robotic ground vehicle with different sensory systems;

[0028] Fig. 3: Shows the working algorithm for the robotic ground vehicle;

[0029] Fig. 4: Shows the types of obstacles found in crop field;

[0030] Fig. 5: Shows the architecture of the autonomous ground robotic vehicle;

[0031] Fig. 6: Shows the schematic circuit diagram for the ground robotic vehicle;

[0032] Fig. 7: Shows the different sketch views of the ground robotic vehicle;

[0033] Fig. 8: Shows the IR camera;

[0034] Fig. 9: Shows the ground robotic vehicle wheel arrangement;

[0035] Fig.10: Shows the Prototype model of ground robotic vehicle (a) Front View (b) Back view (c) Right side view and (d) Left side view;

[0036] Fig.11: Shows the Different parts of ground robotic vehicle;

[0037] Fig.12: Shows the Wheel Working Mechanism with the Acting Force;

[0038] Fig.13: Shows the Balanced Wheel System with Mathematical Notations;

DETAILED DESCRIPTION:

[0039] Fig. 1 demonstrates the strategy of path planning for the agricultural ground vehicle in a crop field. The ground vehicle follows position1 from the start point, then follows position 2 from position1 and continues up to the goal point. The GPS system will set all the positions from start to goal point in the crop field. The ground vehicle will monitor the crops and follows the GPS-enabled positions.

[0040] In Fig. 2, an autonomous algorithm structure with the Raspberry Pi processor and other sensory systems of the agricultural ground vehicle is shown. In this algorithm, we can see the different parameters of the sensory systems are calibrated with the working algorithm. Fig. 3 demonstrates the flowchart of the working algorithm for the agricultural ground vehicle. In the flowchart, the object detection mechanism from the live images is shown with the help of a defined threshold value. Also, the path following strategy is shown in the flowchart. In Fig. 4, the black and white images of obstacles in a crop field are shown.

[0041] Fig. 5 demonstrates the hardware architecture of the autonomous robotic ground vehicle. One rechargeable lithium-ion battery with an input voltage of 12.6V, output voltage of 10.8~12.6V, and capacity of 6.8Ah is the power source in the hardware architecture. The heart of the robot hardware is a Raspberry Pi 4 Model B having a 1.5GHz quad-core Cortex-A72 (ARM v8) 64-bit processor of Broadcom BCM2711, with 4GB RAM. The processor is interfaced with a 5-megapixel OV5647 sensor infrared camera with 1/4inch CCD size, 1.2 Aperture (F), 8mm adjustable focal length, 40-degree field of view, and1080 pixel resolution. An L80 GPS module with an embedded patch antenna is interfaced with the processor. The motor driver circuits are used for direction and speed control of the ground vehicle. 12V DC motors with 43 N-cm torque and 487 RPM speed are used with the motor driver modules. The aluminum wheels (Bush type rollers) with 100 mm outer diameter, 18mm bore (hole) diameter, 54mm wheel thickness, and 340 gm weight are used in the system. 9 rollers arranged in 45 degrees to the plane of the wheel. Relays with 2A maximum DC load current, 24V maximum DC load, and 10 msec operating time are used with the water pump. Also, the sprayer system is attached to the hardware system.

[0042] Fig. 6 demonstrates the schematic circuit diagram for the hardware of Autonomous Ground Robotic Vehicle. As shown in the schematic circuit diagram, the whole system of AGRv is powered by 12-volt Lithium-ion batteries. A couple of discrete components, capacitors and inductors are used at the input and output of regulator IC (UA7805CKCT) to filter the input signal from unwanted spike/noise and to prevent from sudden voltage drop. The 5V regulator IC is used since controller board and other interfacing devices (GPS, Relay) are operated only in 5 volts of supply.
To control the movement of AGRv, a motor driver unit is designed to operate at very high current. Four individual H-bridges is designed using couple of NPN (TIP122-5A) and couple of PNP (TIP147-10A) transistors. Four individual motors are connected to the port U1, U2, U3 and U4. The movement of motors is controlled using dedicated GPIO pins (Control pins) of Raspberry Pi board as shown in Table 1.

Table 1: Dedicated GPIO pins to control the movement of motor
Motor port Number Control pins
U1 GPIO 18
GPIO 23
U2 GPIO 8
GPIO 7
U3 GPIO 1
GPIO 12
U4 GPIO 24
GPIO 25

Table 2: Direction of geared motor wheel
State of the control pin Motor port Motor Direction
GPIO 18 GPIO 23
0 0
U1 No Movement
1 0 Clockwise
0 1 Anti-clockwise
GPIO 8 GPIO 7 Motor port Motor Direction
0 0
U2 No Movement
1 0 Clockwise
0 1 Anti-clockwise
GPIO 1 GPIO 12 Motor port Motor Direction
0 0
U3 No Movement
1 0 Clockwise
0 1 Anti-clockwise
GPIO 24 GPIO 25 Motor port Motor Direction
0 0
U4 No Movement
1 0 Clockwise
0 1 Anti-clockwise

[0043] The clockwise and anticlockwise movement of motors is control as shown in table 2. The movements are control using H-bridge principle. To perform the switching operation in a H-bridge, high current transistors are used as a switch. The base of the all these transistors are connected with 4.7K resistor to provide rated current to trigger the transistor. Across the motors four diodes are used to protect transistors from back EMF.

[0044] Similarly, another H-bridge is designed using BC557 (PNP) and BC547 (NPN) transistors to control the rotation of spray pump. The spray pump is connected to port U6 that control by using the control pin GPIO21 and GPIO20.

[0045] To control the pesticide pump motor a 5V-2A relay is used that control by using a N-Channel MOSFET (IRF540). The switching mechanism of MOSFET is control by GPIO16 pin. The MOSFET is in ON state if state of GPIO 16 pin is high otherwise it will be in OFF state.

[0046] The IR Camera and GPS module is directly connected to Raspberry Pi board. A 15 pin FFC (Flexible Flat Cable) is used to Interface the IR Camera with Raspberry Pi board. The GPS module take power directly from Raspberry Pi board and the transmitter of GPS is connected to receiver pin (GPIO 15) of Raspberry Pi board to receive data serially.
[0047] Fig. 7 demonstrates the top, side and front sketch views of the robotic vehicle. It can be seen that the total length of the ground vehicle is 40 cm, the gap between two legs of the vehicle is 18 cm, and the size of the main robotic body is 30×20×25 cm3. The main body of the AGRv is divided into two parts. The upper part is to store electronic circuitry and the lower part is for pesticide tank. The four motors are supported by a beam to provide rigidity to overall design. In Fig. 8, the schematic view of the IR camera has been shown.

[0048] Fig. 9 demonstrates the wheel system architecture of the robotic ground vehicle. The springs are attached to the wheel system give the ability to absorb shock and vibration. The wheel structure has the ability of speed error control like skidding and slipping. Also, the geometric structure of the wheel system gives the ground vehicle to carry double of its total weight. In Fig.10, the different side views of the real-time autonomous robotic ground vehicle can be seen and in Fig. 11 different parts of AGRV are shown.

[0049] Fig. 12 demonstrates the wheel working mechanism with the acting force. At point A, resolving forces horizontally,
F1 sin???-? F2 sin???=0?
or,F1 sin???= ? F2 sin??
or,F1=F2 (1)

Resolving forces vertically,
F1 cos??+F2 cos????-F=0?
or,F=F1 cos??+F2 cos??

or,F=2F1 cos??
or,F1= F/(2 cos?? ) (2)

Smaller the value of ? exerts small force on the side AB or AC.At point B, a lateral force FL will be there due to force F1. Resolving forces horizontally,
F_X=0
or,F_L-F1 sin??=0
or,F_L=F1 sin?? (3)

Resolving forces vertically,
F_Y=0
or,FG1=F_W+F1 cos??

or,FG1=FG2=F_W+F/(2 cos?? )×cos??
or,FG1=FG2=F_W+F/2 (4)

From the equation (4), it can be concluded that the wheel mechanism of the designed AGRv can absorbed roughly double the weight.

[0050] Fig. 13 demonstrates the calculation of the angle ? in the balanced wheel system. In this figure, for ?PQM,
tan??=(R+X)/R
?or,tan???=1+X/R
or,X=R(tan???-1? ) (5)

For perfect balance of the vehicle,
W=2X+4R Where, R=W/4
or,W=2R(tan???-1? )+4R
or,W-4R=2R(tan???-1? )
tan??=(W-4R)/2R+1
tan??=(W-2R)/2R
?=tan^(-1)??(W-2R)/2R? (6)

The angle ? is calculated for the invented AGRv using the equation (6) for perfect body balance.
In case of shock and vibration control, the springs used with the wheel system play a major role. We know,
Spring rate (k) =(d^4 G)/(8D^3 N)
Here, d = diameter of the spring wire;
G = modulus of the torsion of the spring;
D = the mean diameter of the spring coil;
N = number of active spring coils;
For stainless steel, G = 11.2 × 106 psi,
d = 5 mm, outer diameter = 38 mm;
Number of active coil = 4;
So, Mean diameter = outer diameter – internal diameter = 38 – 5 = 33 mm

Now,
G = (11.2 × ?10?^6 pound)/(inch)^2 = (11.2 × ?10?^6× (0.453592)kg)/((25.4)^2 ?mm?^2 )= 7.874×103 kg/mm2
Therefore,
k = ((5?mm)?^4× 7.874×?10?^3 kg/ mm^2)/(8×?(33mm)?^3×4)= 4.279 kg/mm = 42.79 N/mm

Maximum displacement of the spring = 40 mm

So, from hooks law, k =(Force affecting the spring(F_S ))/(Displacement of the spring (x) )

Therefore, Fs = -kx = -42.7×40 N = -1708 N

So, the maximum load force that can withstand by the spring is, 1708 N. Therefore, the spring can support 170.8 kg.

RESULTS;

[0051] The proposed ground vehicle with the working algorithm has been tested in a crop field. The top results are chosen from the unlimited number of experiments. The success rates and accuracy percentages calculated in all the experiments are satisfactory. Because of the GPS updating errors, some failures are also there in the experimental process. The ground vehicle with full load and no load are tested in the crop field. The top five performances of the ground vehicle in a crop field have been shown in Table 3 and Table 4.

Table 3: Real-time Results of the Ground Vehicle with Full Load (17 kgs)
Observation Path Length
(meter) Travel Time
(seconds) Success
Rate (%) Accuracy
(%)
1. 274.554 2165 100 60
2. 278.732 2227 100 60
3. 279.883 2241 50 40
4. 282.641 2288 50 40
5. 284.122 2316 50 40

Table 4: Real-time Results of the Ground Vehicle with Zero Load
Observation Path Length
(meter) Travel Time
(seconds) Success
Rate (%) Accuracy
(%)
1. 266.610 1656 100 100
2. 268.155 1712 100 100
3. 268.475 1774 100 90
4. 269.091 1825 100 80
5. 269.104 1842 100 80

CONCLUSION
[0052] An agricultural ground vehicle is designed with a visual and navigational sensory system for self-position estimation, path following, obstacle detection, insect detection, and pesticide spraying. The path planning strategy for the AGRv is designed in such a manner that the AGRv can visit every necessary location of the crop field. The working algorithm for the AGRv is worked in different terrains with multiple operational tasks. The robotic ground vehicle can carry double of its weight in the visiting crop field because of its balancing wheel system mechanism. Also, the ground vehicle can absorb shock and vibrations because of two springs attached to the legs of the ground vehicle. The ground vehicle with a capable power source has visited the all directed positions in desired time. The average time taken by the AGRv to cover a crop field of approx. 280 meters with all type of scouting and spraying operations is 37 minutes 45 seconds. The proposed robotic vehicle can successfully scout and perform spraying operations with the help of the working algorithm in various agricultural terrains.
REFERENCES`
Dolly Y. Wu and Timothy A. Deutsch, “Monitoring and Control Implement for Crop Improvement”, US 10,721,859 B2, 2020.
Steve Tanner, Aurélien Demaurex and Gabriele Mondada,“Robot Vehicle and Method Using a Robot for an Automatic Treatment of Vegetable Organisms”, US 10,681,905 B2, 2020.
Steven R. Gerrish,“Autonomous Agricultural Robot (AGBOT) for Decision Making and Courses of Action Considering Real-Time Conditions”, US 2019/0133024 A1, 2019.
David Curtis Crinklaw, Chase Schapansky, Richard Vaccari, Jeremy Bischel and Ryan Johnson, “Robotic Agricultural System and Method”, US 9,877, 470 B2, 2018.
Chad A. Ackerman, Scott A. Stephens, Dolly Y. Wu, Mark A. Harris and David S. Mercer, “Scouting Systems”, US 2016/0157414 A1, 2016.
Harvey Koselka and Bret Wallach, “Agricultural Robot System and Method”, US 7,854,108 B2, 2010.

WE CLAIM:

1. An Autonomous Ground Robotic Vehicle comprising:
a GPS sensor to navigate through the planned locations;
a geomagnetic direction sensor for detecting a direction of a vehicle body;
a GPS receiver for recognizing real time travelling position of the vehicle body;
an infra-red camera and ultrasound sensor configured to detect and record real-time terrain obstacles;
a wheel architecture system for perfect balance on odd and even surfaces; .and .
a programmed processor connected with infrared camera module and coupled with L80 GPS module with an embedded patch antenna enabling said vehicle and causing vehicular wheel centre position and axle to follow a target terrain of said vehicle body.

2. The Autonomous Ground Robotic Vehicle as claimed in claim 1, wherein the said GPS sensor is configured to locate twelve different positions from start point to goal point.

3. The Autonomous Ground Robotic Vehicle as claimed in claim 1, wherein said camera is configured to real time detect agricultural field based obstacles and insects.

4. The Autonomous Ground Robotic Vehicle as claimed in claim 1, wherein the camera is configured with adjustable with 45° vertical view and 180° horizontal view with the help of a movable dc motor.

5. The Autonomous Ground Robotic Vehicle as claimed in claim 1, wherein the said wheel architecture is configured to support the robot body to move with the double weight of the total robot body.

6. The Autonomous Ground Robotic Vehicle as claimed in claim 1, wherein the navigation is carried out with reference to specific threshold value for object detection and maneuvre and steer the wheel architecture. to bypass the obstacle.

7. The Autonomous Ground Robotic Vehicle as claimed in claim 1, wherein the said vehicle is configured as a pesticide sprayer that can spray up to a 1.5 ft range.