The IEEE R5 robotics competition 2016 was held in Kansas city. The competition was a find and rescue mission. Victims would be placed in known locations. They would have a certain color and would have to be dropped off at known locations based on their color. A total of 4 victims had to be moved to the right location based on their color. The 4 victims were either yellow or red and had to be moved to the yellow or the red boxes, one of the two locations. The competition rules can be found here:
Competition Details
The theoretical drawing of the playing field was:
This was the most ambitious project for me. Raspberry pi 2 was used as the master to all the devices.The initial design was to build a swarm of robots and so the spacing was an issue. The whole swarm would have to fit within a 1 foot square area and so the design had physical constraint in terms of size. A single camera was used as a sensor and OpenCV was used for navigation. All of the corrections made by the robot was based on the input from the camera. Bascially, angles from the lines of the walls were used to evaluate the which direction the robot had to move in order to properly align. The robot also used a four bar mechanism to grab the victims. Since everything was closely packed, the stepper motors had minimal distance between them and so heat generated by the motors was a massive issue. Custom heat sinks were built from copper plates. We had to bend the copper plate and create fins in order to get the heat out of the stepper motors. Overall the robot was an ambitious project but failed to deliver because it was too complicated. The swarm was never built and only a single robot was used during the competition. However, the robot was a nice achievement and could have been used with Aritificial Intelligence (AI).
Below is a .gif file which shows how the mechanism picks up any object. The four bar mechanism was first designed on SolidWorks and then built for the robot.
The video shows a run for the competition and after delivering to the red box the servo motor never fell down. This was a weird issue for us and this affected our alignment.
Competition Details
The theoretical drawing of the playing field was:
 |
| Competition field with dimensions in inches |
This was the most ambitious project for me. Raspberry pi 2 was used as the master to all the devices.The initial design was to build a swarm of robots and so the spacing was an issue. The whole swarm would have to fit within a 1 foot square area and so the design had physical constraint in terms of size. A single camera was used as a sensor and OpenCV was used for navigation. All of the corrections made by the robot was based on the input from the camera. Bascially, angles from the lines of the walls were used to evaluate the which direction the robot had to move in order to properly align. The robot also used a four bar mechanism to grab the victims. Since everything was closely packed, the stepper motors had minimal distance between them and so heat generated by the motors was a massive issue. Custom heat sinks were built from copper plates. We had to bend the copper plate and create fins in order to get the heat out of the stepper motors. Overall the robot was an ambitious project but failed to deliver because it was too complicated. The swarm was never built and only a single robot was used during the competition. However, the robot was a nice achievement and could have been used with Aritificial Intelligence (AI).
Below is a .gif file which shows how the mechanism picks up any object. The four bar mechanism was first designed on SolidWorks and then built for the robot.
 |
| Four Bar Mechanism Design on Solidworks |
The video shows a run for the competition and after delivering to the red box the servo motor never fell down. This was a weird issue for us and this affected our alignment.
Comments
Post a Comment