Group items tagged

The Nv neuron circuit may look familiar as it is found in most CMOS data handbooks as an example of a simple edge detector or one-shot or mono-stable circuit application.

robot learning

variable delay

short-term memory

The motor is driven in either the forward or reverse direction, but will swap polarity if the motor encounters too sudden or great of a load

By just tying a neuron's bias resistor to Vcc, rather than to ground, you can make a "regular logic" (active high) Nv:

putting diodes and other resistors in parallel to give different charge vs. discharge rates

>

One essential difference is that the Nv responds immediately to an input, and sends the output for a time duration -- the delay occurs AFTER the output is sent. The Nu responds to an input after a delay and sends the output continuously -- the delay occurs BEFORE the output is sent.

"on" first, then a delay, then "off"

delay, then "on", stays "on"

Behavior-based robotics is quite simply the design of robots where there are no internal "models" of the environment. Instead, the robot's action is state-machine driven via inputs gleaned from the robot's sensors.

The robots do not build a model of their world they simply act in response to the things they encounter whilst existing there.

This form of robotics has proved to be successful in environments that are unknown to the robot, environments that are busy or noisy such as a place with moving objects or people

An important part of the behaviour based theory is "embodiment" This means that a robot must be embodied, have a presence (it is an entity in itself).  In order to react the robot must be surrounded by the real world.

We assume that the slowly changing waveform at the Nv input produces a single output transition when it crosses the Nv switching threshold. This is accomplished by using a 74HC14 or similar CMOS Schmitt trigger. When 74HC04 or 74HC240s are used for Nv inverters some nasty oscillations occur during switching which makes these inverters unsuitable for most Nv networks greater than 2 Nv in series.

Note that behaviors themselves can have state, and can form representations when networked together. Thus, unlike reactive systems, behavior-based systems are not limited in their expressive and learning capabilities.

The basis behavior set of a system provides elements that are not further reducible to each other and that, when composed by sequential or concurrent execution, produce the complete behavior repertoire for the system.

Such basis behaviors are constructed, learned, or evolved from stable, robust interaction dynamics between the agent/robot and its environment, and serve as a substrate for the system's more complex behaviors.

My work generalized the notion of basis behaviors to multi-robot interactions, and demonstrated how a small set of basis behaviors per robot can be used to demonstrate a rich repertoire of individual and group-level behaviors, including following, flocking, homing, herding, aggregation, dispersion, and formations.

multi-robot learning

learning by imitation

uncertainty, non-stationarity, and real-time constraints