Maximizing Science Return and Small Satellite Capabilities with Model-Based Transmission Reduction

Distance from Earth, antenna size and power limit the data rate possible from a spacecraft to Earth-based controllers. For small satellites, particularly those beyond Earth orbit, all three are significant limitation factors. Small satellites are limited in the size of the onboard antennas that they can support and the size of solar panels that can be utilized to generate power. Moreover, most small satellite programs cannot afford access to extremely-high-gain antennas, such as those in NASA's Deep Space Network. Model-based transmission reduction (MBTR) provides a solution for this, by performing onboard processing to assess the value of the data (based on its relevance to a model shared a priori between the Earth-based control station and the spacecraft), the amount of data that must be transmitted to convey a given scientific finding is dramatically reduced. This is critical for deep space missions and can also provide a productivity boost for Earth-orbiting ones. The four levels of MBTR are presented. Each level is progressively further removed from raw data transmission and consequentially the science/knowledge value of each byte of data transmitted is significantly greater. At the first level, model based data transmission (MBDT), data is prioritized for transmission based on its impact value on the a priori model. With model-based data analysis (MBDA), context-aware analysis is performed to determine what areas are the most important (as opposed to just the most different from the model, like under MBDT). The third level, model-based result transmission (MBRT) produces a scientific conclusion based on onboard analysis. Finally, with model-based findings transmission (MBFT), the highest level, the model is updated to reflect the collected data. All levels of MBTR transmit low-level validation data upstream along with their high-value data stream. This allows validation of the performance of the spacecraft and the onboard MBTR software. Each level requires progressively more onboard computational power to perform the work. Experimental results with using MBDT for image data transmission are presented. Initial results for processing this data using MBDA are also discussed. Transmission requirements (e.g., bytes of data required to convey a given piece of knowledge) and accuracy are presented as key metrics for MBTR evaluation and the presented examples are evaluated utilizing these metrics.