Automated Charging Dock 🔋

In collaboration with YITU Technology, my team designed an automated charging point, aimed at reducing human intervention.

I was in charge of ideating, 3D printing the mounts for the robot and programming the line-tracking portions.

Proof of Concept

We were initially inspired by the air-air refuelling mechanism used by planes and toyed around with the idea of using a cone-shaped funnel to charge the robot.

Hence, we played around with the idea of using a magnetic charging cable to connect the charger to the robot as the “locking” mechanism. The video on the right shows our initial testing of the placement and tolerance of the magnetic cable.

Fine-tuning

What we discovered was that even though there was some room for error, a moveable platform was necessary to compensate for any inaccuracies in the robot’s navigation, as well as it’s lack of maneuverability in the lateral axis.

To allow for fully autonomous operation, the dock is designed with an IR sensor which serves 2 purposes : 1. To detect the movement of an incoming robot into the dock, 2. Fine adjustment of charger position to successfully attach magnetic port.

If you look closely, you’ll also notice that the moving platform was salvaged from an old 3D-printer 🤭.

Putting it all together

To incorporate it all together, we worked on the navigation portion of the robot.

We worked on 2 forms of navigation. First of all, we used the SLAM navigation package on the robot, utilizing the LIDAR. While it was useful for navigating to rough positions, it was not accurate enough to move towards the charging unit. After doing some research, we felt that line-tracking was the most cost-efficient, precise and least complex method for the last-mile navigation. Hence, we installed 5 IR sensors on the front of the robot. We then made a subscriber script that listens to the IR sensor readings and decides motor actions through publishing on the “cmd_vel” topic.

The video on the right shows the line-tracking in action.

The overall charging operation is as follows:

The “battery_state” topic publishes voltage levels below 11.0V (low battery)
The robot moves into the dock autonomously (through LIDAR & last-mile line tracking)
IR light gets reflected by the incoming robot with a reflectance strip attached.
IR sensor detects increase in reflected IR light.
Arduino performs averaging of IR reflectance values to verify presence of robot.
Upon confirmation of the robot in the charging dock, Arduino uses a stepper motor to move the charging dock.