Waypoint following and SONAR based mapping in�Unmanned Underwater Vehicle using a novel�control law

BTP II�~Shantanu Shukla �190100107

Objective and Working Scenario

• To use the control law proposed in BTP I for following a given set of waypoints underwater
• While moving along the waypoints, simultaneously map the seabed and detect the topography using a multibeam SONAR installed on the vehicle

​

Vehicle : RexROV �Simulation Environment : Gazebo�

2

3

Control Law for Homing

However, commands needed for RexROV - Thrust and Torque commands with respect to the vehicle frame

- Given through the thrusters present on the vehicle (8 thrusters) �- Thrust in a given direction (x, y, z) �- Torque about a given axis (x, y, z)

4

Homing Results

(a) - shows a relatively straight path towards the origin with less overshoots �(b) - shows a more winding path towards the origin with higher overshoots

• Perform the homing maneuver for successive waypoints through shift-of-origin
• With no Thrust/Torque commands - vehicle experiences an upward force due to the ocean. This is countered by an additional constant downward force to reduce the pitch error.

​

• Set of waypoints to be followed : (10,10,-50), (-10,10,-50), �(-10,-10,-50), (10,-10,-50), (10,10,-50) - square
• Once we are within 1m of a waypoint, we shift the origin to the next waypoint and perform homing with respect to it.

​

5

Waypoint Following

6

Waypoint Following

7

Waypoint Following

Isometric view of the Path followed for 5 sets of values for the 4 parameters. �(e) is the best path taken with relatively low deviation from the desired path

(a)

(e)

(b)

(c)

(d)

8

SONAR based mapping

SONAR : multibeam-p900 (available in uuv-simulator)

(c)

(a)

(b)

Rotation + Translation

While the vehicle follows a set of waypoints - perform the above computation for every detection and store it - export - interpolate - plot (done in MATLAB)

9

SONAR based mapping - Results

Overall topography is determined correctly

�Several spikes in the mapped surface - reduced precision.

10

SONAR based mapping - Results

Overall topography is determined correctly �Several spikes in the mapped surface - reduced precision. ��Plain surface - similar observation of spikes in the mapped region�Possible cause - The oscillations along the pitch

11

SONAR based mapping - Analysis

Run 2 cases - �(1) Vehicle left on it’s own with a random initial pitch at (10, 10, -50) �(2) Vehicle left on it’s own with a 0 initial pitch at (10, 10, -50)

(1)

(2)

Observation - in case (1) the peaks of the pitch of the vehicle correspond to the peaks in the mapped height (z)�In case (2) with negligible pitch (<0.05) - no oscillations are observed in the mapped height (z) �Thickeness of the band of height observed - sensor noise. �Oscillations in the observed height - correlation with vehicle pitch

12

SONAR based mapping - Analysis

Similar Results observed for scans performed while waypoint following -

Improved Scanning - �Set the initial pitch to be 0 �Instead of following waypoints we set the vehicle right above a Wall Block�Give it a constant yaw and calculate the co-ordinates of the scanned points.

13

SONAR based mapping - Wall Block

• Although we get a topographical idea, elimination of sensor noise and pitch correlation in the calculation
• Simultaneous calculation implies - by the time the co-ordinates of a particular point are calculated the actual pitch might change leading to errors
• Modifications and tuning to include dynamic obstacle avoidance and more accurate waypoint following

14

SONAR based mapping - Conclusion & Future Scope

THANK YOU

15