Hello FORUM,
I’ve been stuck on an idea and can’t seem to shake it. What I want to do is build a robot that would simulate a defender (for basketball and other sporting applications) and basically stay in front of a human by tracking its movement but also follow the human should it move out of a preset range. What sort of sensors should I use? I feel as though I’m constantly swaying between different designs and not sure which to stick with. Vision-based using the kinect or pixycam? An array of different sensors (ultrasonic, IR, etc.,)
Thanks so much for your input and help!
Cheers,
Jesse Levi
A project of this scope will need to be split into many segments, and a lot of research and development will be needed.
How much robotics experience do you have, and what kind of budget are you prepared to invest?
Some preliminary issues you will encounter:
]Creating a robot tall enough to block a shot, but stable/:m]
]Creating a robot that can resist a high force so far away from the base without falling over/:m]
]Tracking a human who is moving quickly/:m]
]Determining how to move around the human without causing harm/:m]
You might try really small and create a “goalie” robot which starts by tracking a colored ball and blocking it.
I’m quite certain I would be considered a novice by any standard, especially yours.
As far as investment is concerned, I’d rather keep it as low budget as possible. Not because I’m opposed to spending money but rather because I would not want to create something that the very people I imagine getting the most out of such a creation are unable to afford it.
The robot would not need to block a shot or steal a ball from the player. Heck, it could just be a moving base with a telescoping pole in the middle with a foam or plastic mannequin that has articulated arms attached to it. The top would not need to move, simply provide a “barrier.” Most basketball players will use chairs, cones, or even a ladder to help simulate shooting over a defender or dribbling around a defender and they see improvement in their game because of it. How much more so with an agile object that can be changed to different heights?
I would love to build something more advanced but I realize how costly and advanced that is. I’m simply not ready to attempt such a project on my own. Perhaps in the future.
Thanks so much for your response.
Start simple and work up. A “simple” approach would be to use a single sensor and a single actuator. The actautor responds to the input from the sensor. You can create a simple “goalie” this way. The next step would be to try a vision sensor mounted on a pan/tilt and have it track an object. This will give you a ton of valuable experience and not break the bank. Physically larger projects almost always require large investments.
Thank you! I will take your advice and start simple. Will start a new post once I begin building in the appropriate forum.
