IDR is a complete software package consisting of state-of-the-art machine learning algorithms to reduce complex sensor bias. `Bias` is defined as the delta between the sensor’s measurement and the true value. Bias can come from sources other than the raw sensor, programmatic errors and platform updates can cause errors not modeled by simulation. IDR models these complex interactions by learning from real world data. Additionally, IDR is evolvable and can quickly generate a new model for major platform updates once data is available. Currently used to analyze the APY-9 radar, IDR is platform agnostic and is applicable to any sensor platform.


  • IDR eliminates the process of statistical whack-a-mole by replacing static models based on simulation, with an adaptable and evolvable model based on real world sensor data.

  • Using machine learning, complex sensor bias can be modeled and reduced

  • IDR’s “Man-in-the-middle” design prevents modifying downstream processes

  • Platform agnostic and fast, operating in real-time with minimal hardware requirements