Problem Definition

Following the start of the summer program, I was given free reign to design the navigation system for autonomously docking CubeSats together. It was desired that the CubeSats would have minimal communication with one another, and for them to be able to find each other at long range.

First the problem was simplified to incorporate only two CubeSats; one stationary, acting as a docking Target, and the other a free moving Pursuer that docks onto the Target. Navigation would also only be concerned in 2-D space (a tabletop), rather than 3-D space or even an orbit. In addition, obstacles were not considered, and it was assumed the movement and docking system listened perfectly to the brains of the operation: the navigation system. These modifications of the problem allowed for the basics of navigation to be strung out before the more complex movement and tracking systems would have to be involved. Nevertheless, this still left a fairly complicated problem of having one CubeSat autonomously find the location of another at range and then approach, align, and impact dock with the Target's docking face.

After a few weeks of research, I devised a four step system for having the CubeSats find and navigate towards each other using radios, a camera coupled with computer vision, distance sensors, and lasers.

  Radio  Distance Sensors     Camera         Lasers

1. Radio Direction Finding: 

I wanted to have a reliable system that for minimal interaction between Target and Pursuer. Radio signals allow for this as they can be directionally traced based off their signal strength. By equipping the Target with a radio transmitter, and the Pursuer with a directional radio receiver, the Pursuer could find the direction of the Target relative to itself, by moving its directional radio in a circle, and noting which direction had the strongest signal. This would denote the direction of the Target, as the strongest would be coming from the Target's location.

2. Approach:

After finding the Target's direction, the Pursuer would need to come within a close enough range, so that its on board camera could look for and find the docking face. Thus, the Pursuer would move forward until the camera began to see Target in its view, no matter the face. It would then use a computer vision algorithm to center the Target in view to continue its approach. Upon the distance sensors detecting the Target within a certain proximity, and the Target is still centered, the Pursuer begins a circling maneuver around the Target until its docking face comes into view. 

3. Docking Face Alignment:

When the docking face comes into view, the Pursuer would stop its circling and begin strafing left or right towards the Target. Using a combination of lasers, distance sensors, computer vision, the Pursuer would center the docking face and rotate to align with it. Two lasers come out from the Target and are intercepted by the Pursuer's laser receivers. Due to the lasers' narrow beam width and straight path, they are a strong means of helping the Pursuer align with the Target, if it can receive a laser to guide it. The distance sensors alert the Pursuer about its rotational position relative to the Pursuer's docking face. If one sensor is reading a longer distance than the other when the Pursuer is facing the docking face of the Target, then the Pursuer knows that it needs to rotate to even those distances. The camera and computer vision convey where the Target is relative to the Pursuer, which way the Pursuer needs to strafe, and if the docking face is centered. 

4. Docking:

The last step involves a few checks to make sure the two faces are aligned, and then the Pursuer makes its final move forwards to dock. It checks that its laser receivers are lined up with the lasers on the Target, its distance sensors are reading the same distance to the Target, and the docking face is centered in the camera's view. It then commences docking.

The following table outlines the major components and their functions on the two CubeSats:

Target

Pursuer

Radio Transmitter:  Used for sending radio signals to the Target. Laser Transmitters: A set of lasers used to help guide the Pursue.

Radio Receiver:  Directional receiver that will be used to search for the direction of the Target. Camera: Used to track the Target with the aid of a  computer vision algorithm. Distance Sensors: Used to approach the Target and align with its docking face. Laser Receivers: Used to sense if the Pursuer’s docking face is aligned with the lasers on the docking face of the Target. Raspberry Pi: Brains of the operation to control the camera, run the computer vision algorithm, and read sensor data from the Arduino Nano. Arduino Nano: Used to read sensor data output.