If you put some real thought into the mechanics of actually implementing such a device, you'll realize that your design goals are impossible to achieve.

You hit major problems as soon as you start. You'll need sensors on both wheels to detect speed differentiation. These will need to be hub integrated for an ABS application. Retrofitability and cost effectiveness are instantly gone. You'll need electronically controlled hydraulic brakes. Where's the reliability in that?

I'm also unconvinced about the physics of a bicycle ABS.