QGRAD2™ improves computational efficiency of accurate sensor fusion by 10x over prior technologies
- QGRAD2™ accomplishes precise sensor fusion with a smaller code footprint and fewer instructions per update.
- Fewer computations = improved efficiency. QGRAD is 10x more efficient than existing Kalman filter algorithms.
- The efficiency of QGRAD2™ can be applied to improve update rate by over 300%, or can be tuned to reduce heat and improve battery life in wearable applications.
- QGRAD2™ has a dynamic smoothing process which self-tunes sensors to improve responsiveness and precision under highly dynamic conditions, while still providing stability in static conditions.
How to Get QGRAD2™
Connect the power of QGRAD2™ to your sensors by licensing the patented firmware.
Install our high-performance, ultra-low-profile, QGRAD2™-powered sensors in your product.
A range of specialized and customizable bodies make our sensors the simplest, most versatile way to experience QGRAD2™ performance.
A Revolutionary Advancement in Sensor Fusion
Inertial sensor systems use multiple sensors of different types along with sensor fusion algorithms to compute useful output data. Sensor fusion algorithms utilize the data that comes from sensing elements such as gyroscopic sensors, acceleration sensors, and magnetometer sensors which are each sensitive to different sources of and types of error. Sensor fusion is a process of combining sensory data derived from disparate sources, such as those described above, so that the resulting output information has greater accuracy, less uncertainty, and less sensitivity to error sources than would be possible if the sensors were used individually.
Since sensor fusion is an algorithmic process that combines the component sensor data, different algorithmic approaches may exhibit different performance characteristics. The main performance characteristics of interest are computational complexity and accuracy of the sensor-fused output.
Algorithms with higher computational complexity require the execution of more computations to arrive at a sensor-fused output, which, in turn, can result in slower update-rates, higher latency, and more energy usage. Algorithms with higher accuracy will produce a sensor-fused output that more closely approximates the real-world condition of the sensed object. It is often considered that selecting a sensor fusion algorithm involves a set of possible trade-offs; high-accuracy requires high-complexity and, thus lower update rates, high-speed requires reduced complexity resulting in reduced accuracy, but this need not be the case.
Yost Labs, Inc. has developed the QGRAD™ family of sensor fusion algorithms, along with a suite of related methods, that allow sensor fused outputs to be produced using a method that exhibits both high speed performance and high-accuracy results.