I've played with it, and checked the source code, and this looks like a really, really simple control loop mechanism. It's basically P controller (as in, PID controller with I and D set to zero) and with this simple on/off "mode" selector wrapped around it that basically turns it off sometimes.
The performance seems _really_ bad: set manual input to -0.5 and watch it oscillate like crazy, with a very large amplitude. That mode selector is really not helping here!
The visualization is nice and very high-tech, but a well tuned PID controller would perform much better. Heck, even a simple P controller would likely perform better as well, at the expense of small steady-state oscillation.
hey, thanks for taking a look. it does borrow from control-style behaviour, but it’s tuned for adaptive balance rather than fixed stability. erevos keeps re-weighting its inputs around a shifting baseline. the focus is adaptation and flexibility in decision-making rather than deterministic outcomes. interested to see what kinds of systems this could be used in.
This is a live preview of Erevos v1.3, a minimal adaptive control system.
It detects when its internal processes start to overload and automatically re-balances itself toward baseline.
The demo visualises this feedback loop in real time — you can watch it shift between stress, correction, and stability as inputs change.
I built it as a proof-of-concept for adaptive stability — showing how a simple feedback loop can detect its own stress and correct itself without supervision.
Demo: atlas-systems-labs.com/demo
Code: github.com/Atlas-Systems-Labs/erevos-demo
I’m curious how others see this kind of adaptive feedback loop fitting into areas like simulation, robotics, or behavioural systems.
thanks! erevos is a lightweight adaptive stabilisation loop. it can sit inside control systems, ai behaviour layers, or anywhere you need competing signals to balance rather than fight. it keeps systems steady while still letting them react to change.
What is the system model being simulated here?
I've played with it, and checked the source code, and this looks like a really, really simple control loop mechanism. It's basically P controller (as in, PID controller with I and D set to zero) and with this simple on/off "mode" selector wrapped around it that basically turns it off sometimes.
The performance seems _really_ bad: set manual input to -0.5 and watch it oscillate like crazy, with a very large amplitude. That mode selector is really not helping here!
The visualization is nice and very high-tech, but a well tuned PID controller would perform much better. Heck, even a simple P controller would likely perform better as well, at the expense of small steady-state oscillation.
hey, thanks for taking a look. it does borrow from control-style behaviour, but it’s tuned for adaptive balance rather than fixed stability. erevos keeps re-weighting its inputs around a shifting baseline. the focus is adaptation and flexibility in decision-making rather than deterministic outcomes. interested to see what kinds of systems this could be used in.
This is a live preview of Erevos v1.3, a minimal adaptive control system.
It detects when its internal processes start to overload and automatically re-balances itself toward baseline.
The demo visualises this feedback loop in real time — you can watch it shift between stress, correction, and stability as inputs change.
I built it as a proof-of-concept for adaptive stability — showing how a simple feedback loop can detect its own stress and correct itself without supervision.
Demo: atlas-systems-labs.com/demo
Code: github.com/Atlas-Systems-Labs/erevos-demo
I’m curious how others see this kind of adaptive feedback loop fitting into areas like simulation, robotics, or behavioural systems.
What can this be used for?
thanks! erevos is a lightweight adaptive stabilisation loop. it can sit inside control systems, ai behaviour layers, or anywhere you need competing signals to balance rather than fight. it keeps systems steady while still letting them react to change.