Drishti
दृष्टि/ dṛṣṭi
See the world, and predict what it does next.
Drishti is embodied perception with a predictive world-model — the body's sense of where things are, what they are, and how they will move. It is the model the agent acts inside.
In development
Spatial understanding
Builds and maintains a 3-D sense of the scene from on-board sensing.
Object permanence
Tracks objects through occlusion and motion — they don't vanish when unseen.
Prediction
A world-model that rolls the scene forward so the body can plan against the future.
Intent
Reads the motion of people and machines nearby to act safely with them.
Acting in the physical world requires more than seeing a frame — it requires a model of how the scene will evolve. Drishti maintains that model continuously, on-board.
It gives the agent object permanence and prediction: the cup is still there behind the box; the person is about to cross the path. That foresight is what separates reactive machines from ones that plan.
Drishti is the perceptual substrate Chetana reasons over — the difference between a robot that responds and one that anticipates.
Next system
Kshana क्षण — The edge runtime.
The whole stack, on the robot, in real time.