Nature thinks differently
Brains evolved to solve autonomy in the real world, by default. By using highly differentiated algorithms, they far exceed today’s engineered state-of-the-art (SoTA) approaches.
Opteran has patented these biological algorithms to make machine autonomy as robust and efficient as nature.
Solving vision and perception nature’s way
Opteran “sees” at low resolution, high frame rates stabilized to 3DoF rather than high resolution, low frame rates and 6DoF. As a result, they simplify how we understand motion and increase the robustness of sensor noise.
See in visually challenging environments.
Increase robustness with stabilization.
Significantly reduce compute and sensor cost.
- Res. vs. FPS
- Cost
- Noise
- Latency
- Range
- DoF
Engineered Approach
High resolution and moderate (E.g.20-60fps) frame rate.
Opteran Approach
High frame rate (90-150fps) and low resolution.
Build robust maps dynamically as you move
Opteran creates robust navigation by adding diversity between understanding motion and position, resolving SLAM fragility, and allowing ultra-low memory maps to be built on the edge, at city scale without processing data or training.
Resolve SLAM single point of failure.
Ultra low-memory mapping.
GPS free navigation.
- Memory
- GPS
- Robustness
- Learning
- Privacy
- Network
Engineered Approach
Requires GB of data per meter squared + heavy processing.
Opteran Approach
Unique patented approach to mapping and localization.
104x
More efficient than tradition SLAM
Detect collisions in the real-world at real-time
Opteran avoids collisions by responding dynamically to what it can see, eliminating the delay from managing occupancy maps with costly replanning.
Real-time awareness.
See small static and dynamic objects.
Straight through processing.
- Efficiency
- Dynamic objects
- Sensing
- Aliasing
- Small objects
- Speed
Engineered Approach
SLAM struggles to reconcile moving sensor data points with the occupancy map whilst rejecting outliers.
Opteran Approach
Inherent understanding of moving objects. (E.g.people. other objects)
Nature evolved to be low SWaP$ by default
Opteran algorithms evolved to enable insects to navigate so are innately edge only, small, lightweight and ultra-low powered. Allowing them to be deployed onto the low-end silicon e.g. CPU, FPGA to ASIC.
Lowest power consumption.
Low cost compute.
Minimal CO2.
- Size
- Weight
- Power
- Cost
Engineered Approach
Sensors and compute heavy.
Opteran Approach
Ultra small, low-cost sensors and lowest cost silicon.
