Waypoint Guidance for Unmanned Ground Vehicle
Waypoint guidance and navigation are indispensable fields in the research in autonomous vehicles. Following user-defined paths and seeking goal locations is the main objective in waypoint guidance. In this work, an autonomous Unmanned Ground Vehicle (UGV) is commanded to drive itself along a path de
2025-06-28 16:36:44 - Adil Khan
Waypoint Guidance for Unmanned Ground Vehicle
Project Area of Specialization RoboticsProject SummaryWaypoint guidance and navigation are indispensable fields in the research in autonomous vehicles. Following user-defined paths and seeking goal locations is the main objective in waypoint guidance. In this work, an autonomous Unmanned Ground Vehicle (UGV) is commanded to drive itself along a path defined by a series of waypoints, which will be user-defined through a mapping interface. In the case of autonomous navigation, a UGV usually employs Light Detection and Ranging (LIDAR), cameras, and Global Positioning System (GPS) receivers to avoid obstacles and follow a set of GPS waypoints that define a path for the UGV to cover. It has to autonomously navigate through its surroundings and execute its mission; hence including obstacle detection and avoidance. This research consists of implementing an environment mapping system that will allow the UGV to follow the path created by a set of waypoints and detect potential obstacles using real-time sensor data. Implementation will be done through National Instruments (NI) LabView, with NI myRIO being used as the main processor. Data will be communicated through a wireless network.
Project ObjectivesThe objectives of the project can be summarized as follows:
- Implementation of waypoint guidance through GPS
- Path Navigation
- Autonomous obstacle avoidance
- Sensor integration
- Wireless networking
The project has two parts
- The Pure Pursuit algorithm to follow the series of waypoints
- The Obstacle avoidance, to autonomously navigate the created path
Pure Pursuit Algorithm:
Pure pursuit uses cross-track errors and look-ahead distances (as explained further on) to make the steering as humanly as possible, in terms of measuring the error before time and hence closing the error during driving the vehicle. The cross-track error and the look-ahead distance are not measured using sensory data – they can be – but are mostly calculated through odometrical equations. Many applications of pure pursuit are dependent on GPS and IMU – Global Positioning System, Inertial Measurement Unit. IMU uses gyroscopes to measure yaw, roll, and pitch angles for information on vehicular position; they also provide linear acceleration information. This is found within the myRIO and hence an external IMU module is not used in future UGV runs. The values from the IMU are used with odometrical equations and dead reckoning to correct any inaccuracies from the GPS positioning. They can be made more accurate using Kalman or Extended Kalman Filters.
In addition to this, the pure pursuit controller used below is based on the traditional Ackermann steering model. An advantage to the pure pursuit controller is that they have no restriction on whatever path is set by the front-end user – unless there is a hardware boundary limit set. The paths can be set as continuous curvatures that can be potentially
discretized into piece-wise linear segments.
Obstacle avoidance:
The best-implemented method is using an RPLIDAR, and involves preprocessing to make the method more efficient and requires less processing. The RPLIDAR A2 captures 8000 measurements per second, ranging from 0 degrees to 360 degrees, up to 18 meters. The data recorded is the reflection of the obstacle and contains three parameters: 1) Degrees 2) Distance 3) Signal Strength. Therefore based on the degrees recorded we can classify measurements into 4 different classes. Forward, Left, Right, Back. Each of these classes is divided into two further division’s namely long-range and short-range measurements. Measurements between 0 cm and 110 cm are placed in the SR category while measurements between 111 cm and 1800 cm are placed in the LR category. The SR measurements carry more weight compared to LR ones.
The algorithm drives the vehicle in the direction of the required position until an obstacle is detected. When the vehicle cannot keep going forward due to a detected obstacle, the algorithm then checks left and right to determine the possibility of a turn. If both left and right directions are free of obstacles, the algorithm favors the direction closer to the required position. If every direction is blocked, the algorithm drives the vehicle in the reverse
direction while simultaneously checking for a different possible turn.
The project can be used in a variety of applications, thus introducing a modern way of achieving autonomous control in different industries.
(Since the project involves navigation of path created by a set of user-defined waypoints. it can be used in the following applications)
- Surveillance Missions with Multiple UGVs
- Autonomous crop harvesting
- Self-driving cars
- UAV/UGV Platform for Detection and Fighting
The final deliverables of the project would include:
- A UGV, that can navigate any terrain, has eight motors that will not only allow linear movement but have flippers to climb onto objects or provide more torque when moving in rough terrain.
- A UGV that would be able to follow a path created by a set of waypoints, ideally with no limitation of the operational area.
- A UGV would autonomously detect obstacles within a 100m range, avoid them and navigate back to its original path.
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 72000 | |||
| RPLIDAR | Equipment | 1 | 30000 | 30000 |
| Motor Drivers (Saber-tooth)) | Equipment | 2 | 15000 | 30000 |
| Joystick | Equipment | 1 | 7000 | 7000 |
| Others | Miscellaneous | 1 | 5000 | 5000 |