Design, Development, and Testing of an
Unmanned All-Terrain Vehicle
Brief Documentary
Introduction
With the evolution of technology, the demand on sending robotic devices to risky territories is increasing. In some industries, furnaces are approached by workers, which increases the health concerns. In other fields, operators patrol wells 24/7 in order to collect data, shutdown/switch on wells, or collect samples; tasks which are too easy for humans, and not too complicated for robots.
From another point of view, Bahrain do not have any laws or regulations that govern the method that the handicapped individuals can use to lift themselves to one of the public transportation facilities. Regulations do cover the existence of special transportation that is free of charge, but it is available only for treatment and education purposes. This suggests that the existence of a mechanism that can help in climbing steps and stairs is a necessity.
In this work, I will present the design, development, and testing of a vehicle that can reach a given GPS coordinate to collect data, or samples (depending on the application), and then return to its station. In addition to that, a climbing mechanism is to be designed and implemented in order to assist the vehicle in climbing steep obstacles. This is approached by using SOLIDWORKS to design the mechanical components, and then fabricate them as necessary. Another aspect is to construct the electronic circuits and program two Arduino microcontrollers to fulfill the vehicle’s tasks. The vehicle can successfully navigate automatically, or manually as required by the user. In addition, the climbing mechanism is working as necessary, where the vehicle is able to detect obstacles via ultrasonic proximity sensors, which engages the rack and pinion mechanism, and then climb the object.
Design Conceptualization
A. Climbing Mechanism Development
The design of the Unmanned All-Terrain Vehicle is oriented towards ensuring that a climbing mechanism is allocated on-board. The reason behind this is that the type of the mechanism significantly influences the dimensions and geometry of the vehicle’s structure. Using the minimum amount of torque during the climbing operation is a main consideration. Doing so, will ensure that the power requirements can be met easily. For that, a rack and a pinion mechanism is innovated, as fig. 1 below shows a 3D model of the vehicle.
Figure 1. Final model of the vehicle. Rendered by SOLIDWORKS.
The design requires having 7 motors, 2 of which are for raising the vehicle, while 1 is for sliding the vehicle forward above the obstacle, and the remaining 4 motors are for moving and navigating the UATV. In terms of cost, the torque requirements from the motors is not significant, which is calculated at 1.5 N.m for each motor. Therefore, the purchased motors are not expensive, nor demanding significant power. In terms of manufacturing, the transition from rotary to linear motion requires having some mechanical arrangements which are not significantly complicated, and can be fabricated with the available facilities. The method of lifting the vehicle is fairly stable, and it is adaptable to various heights, without significantly affecting the torque requirements. Fig. 2 below shows how the climbing operation is achieved theoretically, as predicted from the SOLIDWORKS model.
Figure 2. Planned operation of the climbing mechanism. Generated by SOLIDWORKS.
The planned sequence of climbing is as follows:
- The vehicle slowly approaches the obstacle, until it contacts it. Once the contact occurs, the vehicle must stop
- The climbing initiates by driving the rack downwards until the vehicle reaches the end of the obstacle.
- The sliding mechanism is activated, and it drives the vehicle forward above the obstacle.
- The rack must go upward once all the 4 wheels are on top of the obstacle.
- Once the rack is all the way up, the sliding mechanism is activated again to bring the slider back to the original position.
B. Algorithm Development
It is important to understand the requirement of each electronic component, and the feasibility of establishing proper connections between the components. As mentioned earlier; the vehicle must be able to navigate to the desired GPS location, in addition to avoiding or climbing obstacles ahead. Based on those requirements, the appropriate electronics and sensors were obtained. For instance, 3 ultrasonic sensors are used in order to measure the distances ahead, and alert the user of any obstacle in sight. A GPS receiver is used to obtain the vehicle’s current location, while an E-Compass (Magnetometer and Accelerometer Module) is used to obtain direction (heading) readings of the vehicle. Limit switches are used to regulate the motion of the climbing mechanism, and fully automate it. In terms of controlling the motors, 4 motor driver circuits are used, to control both the direction, and the speed of each motor individually. All the components are controlled by an Arduino Mega 2560 Microcontroller, and 2 Li-Ion batteries are used to provide power to the whole vehicle.
The logic of the GPS navigation, is based on two main formulas; the Haversine distance formula, and the heading formula.
The Haversine distance formula uses trigonometric functions in order to calculate the great-circle distances between two points on a sphere from their longitudes and latitudes. The formula is given as follows:
Where, d is the distance between two coordinates,
R is the radius of earth i.e. 6371 km or 3961 miles,
Φ1, Φ2 are latitudes of point 1 and latitude of point 2,
λ1, λ2 are longitude of point 1 and longitude of point 2.
As for the heading formula, it will be used to point the vehicle in the right direction while it is navigating. Essentially, the calculated heading angle should be kept as close as possible to the obtained heading from the E-Compass, in order for the vehicle to reach its destination. For calculating the destination heading between the given coordinates and the target’s coordinates, a forward azimuth formula is being implemented. This formula traces a path along the shortest distance between two coordinates, it is given by the following equation:
Where, h is the heading,
Φ1, Φ2 are latitudes of point 1 and latitude of point 2,
λ1, λ2 are longitude of point 1 and longitude of point 2.
By subtracting the target heading from the destination heading, the sign and the value obtained (which is essentially the error between the two), can then be implemented in the program’s algorithm in order to navigate and control the direction of the vehicle.
Once the ultrasonic proximity sensors detect that an obstacle is in front of the vehicle, the user is requested to choose whether the vehicle should climb the obstacle, or avoid it. If the vehicle is asked to climb it, the mentioned sequence in fig. 2, is implemented. Limit switches are installed in the mechanism to send signals to the microcontroller, and indicate the start and completion of each phase of climbing.
Fabrication of the Vehicle
The vehicle is manufactured with the available facilities, after selecting the appropriate materials. The materials that are involved in the vehicle are Aluminium (Alloy 6061), Steel (AISI 1020), Brass, ABS Plastic, Acrylic, and Delrin. All of which were allocated to the various different components, based on the weight, strength, ease of fabrication, and the available form (whether it was sheets, tubes, solid round bars, or 3D printer filament). The engineering drawings of all the components, along with the relevant dimensions were produced from the 3D models. This was done by ensuring that the designs are simplified as much as possible, a technique known as ‘Design for Manufacturing’. The mechanical components were fabricated and manufactured from raw materials by turning, TIG welding, grinding, bending, and drilling. The remaining mechanical components were 3D printed from ABS Plastic. The reason behind that is the lack of facilities to fabricate the complicated components from other more practical metals or composites. AGMA (American Gear Manufacturing Association) standards were followed to design the two rack and pinion mechanisms, to ensure that the ABS material retains enough strength to withstand the loads that are endured by the UATV during climbing. Once that was confirmed, the components were printed.
As for the electronic components, the connections were established permanently by wiring all the components, and programming the microcontroller.
Performance Analysis
Some of the performance parameters of the UATV will be analysed. For instance, the performance of GPS navigation, in addition to some power analysis to one of the motors. The main outcome that is to be achieved from these analyses is to measure the performance of the vehicle and the factors that affect it from different point of views. This will be beneficial for further research and any future studies that are relevant to this project.
A. Navigation Analysis
The UATV was tested for several times, while logging data of current time and date, current latitude and longitude, destination latitude and longitude, altitude, speed, distance to destination, current heading, target heading, and the temperature. The GPS analysis will only look into the linearity of the navigation path. Fig. 3 below shows the results of one of the trials.
Figure 3. Left image shows satellite plots for the navigation path of the vehicle, and the graph on the right shows the same path, plotted graphically (blue). Linear regression of the path is highlighted in red.
The R2 value, which describes how close the line which is generated from linear regression to the plotted dots, was found to be 0.8683. This suggests that the UATV navigate in a linear line with a percentage of 86.83%. Nevertheless, during navigating through this path, the vehicle went through some very rough surfaces, which deviated it from its destination, and caused the program to carry significant corrections. Eventually, the vehicle did reach the destination, with a final error of only 2.29 meters, which is significantly accurate. Therefore, the linearity of the vehicle’s path is not really of a great importance, when considering that it eventually reaches its destination with a very high accuracy. This value can be improved further, by changing the program’s algorithm to a better algorithm, while including an effective control system that can maintain an acceptable steady-state error, instead of the fluctuations that are occurring currently.
B. Climbing Mechanism Analysis
It was verified that the frictional effects will require one of the climbing motors to produce a torque of around 0.375 N.m in order to actually actuate the mechanism. The coefficient of friction which was assumed to carry out that calculations, is taken while assuming that the contact will be for unlubricated surfaces. In order to understand how lubricating the mechanism can affect the amount of power that is required from the motors, an experiment was carried out. The apparatus includes an Arduino microcontroller, a motor driving circuit, and a power supply. The power supply was used to power both the Arduino and the motor driver circuit. The motor was then connected to the circuit output. Then, while controlling the distance that the sliding rack cover, 12 volts were supplied to the driver circuits, and the sliding rack is allowed to move from the beginning to the end, to cover a distance of around 300mm (which is the same distance that must be covered to slide over the obstacle). During the motion of the rack, the voltage and current were continuously logged, at each distance in the rack. Then, by multiplying the voltage and current (since ), the power can be calculated. The first test was carried out before adding any lubrication to the sliding base, while the second test was carried out after applying light grease to the base. The results were logged, and fig. 4 below shows a comparison of the power consumption in respect to the distance, prior to and after lubricating the components.
Figure 4. Analysis of the power consumption before and after lubricating the climbing mechanism.
The difference is greatly noticeable. Without lubrication, the operation is extremely not smooth. The mechanism’s friction is significant, and as a result of that, the motor draws significant current, and the motor driver circuit becomes exposed to multiple current spikes. These spikes shorten the life of the electronic components, and may lead to their failure.
As for the result with lubrication, the situation is completely changed. The power delivery is now much smoother to the motor, with no sudden spikes and such. This suggests that lubrication can have a significant effect on the performance of the climbing mechanism.
This significance is really noticeable when it comes to numbers. Without lubrication, the maximum power spike reach up to 13.27 Watts. As for the lubricated base, the maximum power drawn at certain point is equal to 6.36 Watts. This indicates a very significant difference in power consumption. The unlubricated base demands more than twice of the energy than the lubricated base. In addition, the behaviour is similar when it comes to the average power consumption. Throughout the sliding of the base, the average power consumption for the unlubricated base is equal to 7.83 Watts, whereas the average power consumption of the lubricated base is only equal to 4.66 Watts. This implies that on average, around 40% of the energy requirement is reduced, just due to lubrication. It is worth mentioning that the total mass of the vehicle is equal to 9.8 kg.
In terms of the endurance of the design, the power is directly proportional to the torque that is transmitted through the gear, whereas . This suggests that when more power is required to move the mechanism, the gear will endure more stresses, up to a point which exceed the design limitation, and can lead to the failure of the gear.
The climbing mechanism of the UATV was tested, and fig. 5 below shows the result of the actual model during climbing one of the steps.
Figure 5. Actual model operation during climbing obstacles.
It can be observed that the operation is almost identical to that in fig. 2 above. This suggests that the vehicle can indeed succeed in climbing steps that have a height between 55mm, and not more than 275mm. The minimum height depends on the height position of the ultrasonic sensor, while the maximum height depends on the length of the rack.
Summary
The UATV is equipped with 7 motors, several sensors, and an Arduino microcontroller. The microcontroller receives feedback from the sensors, and does calculations to achieve two tasks; to navigate from one coordinate to the other via GPS, and to avoid or climb obstacles. The navigation is achieved by calculating the distance and heading to destination, and comparing those to the data that is obtained from the sensors. An accuracy of 3 meters was achieved, which is reasonable, even though the linearity of the vehicle was analysed and found to be following a linear line by around 87%. As for climbing, the vehicle was found to be climbing obstacles that range between 50mm and 275mm, and with lubricating the mechanism, around 40% of the power consumption was reduced, and it was found to be equal to 4.66 Watts, which reasonable when we consider that the mass of the vehicle is equal to 9.8 kg.
Further future modifications are possible. For instance, a visual feedback device and additional sensors can be added in order to increase the data collection capabilities of the UATV, and increase the amount of applications that it can be used in. Furthermore, a physical data collection system can be installed as well, in order to collect samples, or carry out some other physical tasks. The Arduino can be replaced with another industrial microcontroller, which retains more memory capabilities, and higher clock speed. This will provide more opportunities to program and implement advanced control system algorithms, and additional features.
I wish to extend my sincere gratitude to my supervisor Dr. Subramanian Chithabaram, Mechanical Engineering Tutor, for his continuous support throughout the period of the project. Thanks to his help, all the challenges that were faced during the project were overcame. I also extend my thanks to Dr. Christakis Papageorgiou, Dean of EDICT (Engineering, Design, and ICT), Dr. Christina Georgantopoulou, Head of School of Engineering, and Mr. Pradeep Nathoo, Programme manager of Mechanical Engineering, for their encouraging comments and beneficial feedback. At last, I would like to thank the technicians Ramadhan Yousif, and Yousif Jawad for their assistance in fabricating the mechanical components, and troubleshooting the electronic circuits.