Today’s robots are still extremely limited in capability to perceive and understand their environment. Ultimately, this shortfall is an impediment to adoption for tasks everyone knows we want robots to perform.
It’s relatively easy to get robots to manipulate known objects in known locations or to navigate a known path to a well-defined destination, but interacting with dynamic, unknown, and complex man-made environments and manipulating objects along the way is another level of difficulty. Heuristic Labs’ technology allows robots to more accurately sense and understand their environment – a pre-requisite to all higher-level tasks.
Advanced Driver Assistance Systems (ADAS) are the stepping stones to complete self-driving/piloting vehicles. These systems are necessarily built on automated perception from on-board sensor data. Current products rely on aging 2D perception techniques which have no knowledge of size and limited ability to determine range or speed; as a result, false alarms occur or the system fails in unexpected or novel circumstances - when it is needed most! Heuristic Labs’ ability to enable next-generation full 3D perception overcomes these challenges, leading to more capable, more aware, and safer vehicles.
In order for UAVs to become widely used and common objects, several safety concerns need to be addressed. For example - obstacle avoidance, in flight or when choosing a landing location. While many current systems have autopilot features supported by GPS, these systems are still ‘stupid’ in the sense that they will run into any obstacle that blocks their path. Giving the device an ability to see its environment to prevent collision accidents will become a standard safety feature. This is also a prerequisite to full autonomy.