Autopilot path planning and obstacle Avoidance Steering control
The design of ship steering control system involves many concentrations in the design algorithm because the dynamics of ships are not fixed and involved nonlinearities in the system. Hence it will not only hold the potential for reducing propulsion losses due to steering, but also maintain tight con
2025-06-28 16:30:34 - Adil Khan
Autopilot path planning and obstacle Avoidance Steering control
Project Area of Specialization Mechatronics EngineeringProject SummaryThe design of ship steering control system involves many concentrations in the design algorithm because the dynamics of ships are not fixed and involved nonlinearities in the system. Hence it will not only hold the potential for reducing propulsion losses due to steering, but also maintain tight control when operating in confined waterways. Almost all conventional control algorithm have been developed but still there is divine need of a robust control system that can perform well for automatic ship steering with negligible error in the ship heading motion. SMC (Sliding Motion Control) algorithm for Autopilot is already been developed without any consideration on obstacle avoidance & climate forecasting information. In addition, finding obstacle avoidance path is complicated that a ship must identify and bypass this with an automated decision making system. We applied a robust data association based path tracking system called integrated probabilistic data association (IPDA) algorithm for tracking the path of desired trajectory. The proposed algorithm initializes tracks for path planning and trajectory referencing using the sensor measurements. The Extended Kalman filter method EKF algorithm predicts the trajectory state and update the track to obtain state estimate. A track quality measure called track existence probability is utilize for false track discrimination (FTD) performance. The FTD confirms the true and eliminates the false tracks. The results are verified using simulation.
Project Objectives•To develop a robust integrated data association based path planning and tracking system called integrated probabilistic data association (IPDA) algorithm for Unmanned Autopilot Ship
•To initialize tracks for path planning and trajectory referencing using the sensor measurements.
•To track the path of desired trajectory avoiding obstacles and false alarm in the cluttered environment .
•To apply Extended Kalman Filter algorithm that predicts the trajectory state and update the track to obtain state estimate of the ship.
•To obtain a track quality measure called track existence probability (PTE) which is used to maintain and update the track automatically.
•To utilize PTE as the performance metric for false track discrimination (FTD). The FTD confirms the true and eliminates the false tracks.
Project Implementation MethodThe proposed algorithm Initializes tracks for path planning and trajectory referencing using the sensor measurements. Path planning block consist of track initialization, IPDA, track update and true track test condition. Tracks are initialized and updated using the measurements obtained in each scan. Track initialization in a cluttered environment results in both true tracks and false tracks. True tracks always follow the target measurements,whereas false tracks do not follow the target measurements. A technique called false track discrimination (FTD) is used to confirm the true tracks and terminate the false tracks. Integrated probabilistic data association (IPDA) it Introduces recursive formulae for data association and employs the probability of target existence as a track quality measure. The practical considerations for IPDA are discussed in this project. If the target trajectory state at scan k is denoted by xk, then the target trajectory propagates by 
Where Fx is the forward state propagation matrix and is the zero-mean white Gaussian plant noise sequence with known covariance matrix Q. The updated probability of target existence is used to confirm and terminate tracks. tracks are confirmed when the updated probability of target existence exceeds a predetermined confirmation threshold and terminated when it falls below a predetermined termination threshold.
The proposed unmanned autopilot ship based on integrated probabilistic data association (UAS-IPDA) algorithm developed the path for desired trajectory avoiding obstacles and obtains the best FTD performance of the trajectory with improved estimation accuracy in the cluttered environment to avoid collision track obstacle Avoiding Path that is Controled. The combination of greater flexibility, lower capital and lower operating costs for this Research based algorithm are key benefit of this project.We will further uses it as for Tracking subject to the Condition.
Technical Details of Final DeliverableThe Extended Kalman filter (EKF) method predicts the trajectory state and update the track to obtain state estimate. Path planner and tracking procedure is applied the ship model. The ship dynamic non-lineaarty due to false alaram and sea waves can be tackled by static clutter reduction and constant false alarm rate method.Static clutter reduction method using KF tracks the signals from static clutter and extracts signals from moving targets in each radar scan in (1). The signals from stationary objects such as background walls generate stationary amplitude and propagation delay
State and measurement equations can be written as (2). ,
where 
is static clutter signal, i.e., = rc(k)[m], ZkCR represents the radar scan in scan k, i.e., 
= r(k)[m], A = , H = 
, and is identity matrix.
The proposed unmanned autopilot ship based on integrated probabilistic data association (UAS-IPDA) algorithm developed the path for desired trajectory avoiding obstacles and obtains the best FTD performance of the trajectory with improved estimation accuracy in the cluttered environment
Check Constant False Alarm Rate (CFAR) Detection in Autopilot
Radar samples with larger amplitude than that of reflected from various backgrounds should be figured out for extracting target signals. In practice, the signal strength varies with respect to time. Therefore, a certain threshold should be used to avoid data loss or miss detection. The threshold can be determined adaptively to the signal using CFAR detection. CFAR detection is a general adaptive algorithm in radar fields to detect the target returns against background noises and clutters . Iparticular, Cell Averaging CFAR (CA-CFAR) detection is applied to the static clutter reduced signals expressed in (2).CA-CFAR detection utilizes a sliding window composed of three cells: reference cell (RC), guard cell (GC), and cell under test (CUT) as shown in Fig. . RCs are placed on the most front and rear of the window. In windowing, the signals in RCs are treated as noise signals. CUT is placed on the center of the window. Signals in CUT are compared with the average of RCs as in (5). If the signal strength of CUT is larger than the threshold which is determined from signals of RC, then the samples in CUT are treated as target signals. The signals processing from each cell can be expressed in (3) to (5) and shown in . GC is placed between RC and CUT to separate the two cells.
where Pn is the average noise amplitude, N, NG, and NCUT are the number of samples in RC, GC, and CUT respectively, and NC is the sum of 2NG and NCUT. X indicates an amplitude of the signal. PFA is a desired false alarm rate, and d is a Boolean index of detection.
We have performed Simulation Based results but only Hardware implementation is required so we are currently work on it.
Final Deliverable of the Project HW/SW integrated systemCore Industry EducationOther Industries Manufacturing , Transportation , Others Core Technology Artificial Intelligence(AI)Other Technologies NeuroTechSustainable Development Goals Good Health and Well-Being for People, Partnerships to achieve the GoalRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 68360 | |||
| Lidar Range Finder Sensor Module | Equipment | 1 | 20000 | 20000 |
| Pixhawk 2.4.8 GPS telemetry complete kit | Equipment | 1 | 25000 | 25000 |
| Raspberry PI 4 Model B 2GB RAM | Equipment | 1 | 12000 | 12000 |
| Steering Auto-Actuator | Equipment | 1 | 5000 | 5000 |
| IR obstacle avoidance sensor | Miscellaneous | 3 | 120 | 360 |
| Ship Model prototype | Miscellaneous | 1 | 5000 | 5000 |
| Ship floating tub | Miscellaneous | 1 | 1000 | 1000 |