If data had mass, the Earth would be a black hole”
– (Marsland, 2015)
Careful data analysis is becoming increasingly important to an organization’s growth in today’s technologically-advanced world. The aim of data analysis is to gain useful insights by converting data into information, in other words attaching meaning to raw data. Computer algorithms are great at pattern-matching and detecting relationships within data, but humans are still an integral part of the pipeline to turn data into useful information. With rapid strides being made in the development of AI algorithms, the human input required might be drastically reduced in the future, however, currently both human and machine are key to the process of making sense out of data.
By modelling and attaching meaning to data, new, unexpected patterns often emerge. The extraction of useful information from extensive databases & archives of data enables the user to make critical decisions for all sorts of applications and fields (Foslien, Guralnik, Haigh; 2004).
The problem of big data in space
The space sector generates massive archives of data for a variety of applications. The tremendous growth of space-related data over the last decade, projected to continue accelerating, has led the community to seek ways to tackle “Space Big Data” (SBD). There are several definitions of SBD leading to a lot of discussion within the community (International Space University; 2016). Nevertheless, in a general sense, SBD is defined to be at the junction of the space industry and the big data world, and it “encompasses all data gathered through activities that utilize space assets” (Marchetti, Soille, Bruzzone; 2016).
The development of innovative technologies for space applications requires advanced use of spacecraft data for a range of applications, including automated system health surveillance, end-of-life management and autonomous operations (Iverson; 2008). In the case of automated system health surveillance, data processing can highlight the presence of outliers in system information; often a signal of malfunctioning subsystem components. The rapid increase in the volume of telemetry data can be traced back to the ever-rising number of sensors being employed on board of satellites, the frequency at which data are being sampled and, last but not least, the number of satellites being launched. The rise of mega constellations for satellite communications (Guidotti; 2017) in particular, has direct consequences on the volume, velocity, and variety of data being generated.
Hence, satellite operators are increasingly faced with making sense of SBD stored in large telemetry databases. Most commercial satellites in Earth orbit provide a regular stream of telemetry data to ground stations, placing the emphasis on developing robust tools for storage, processing, and analysis. As a result, there is increased focus on advanced data mining methods to make sense of all the telemetry data coming from the satellite. Critical mission phases, for instance during launch, require high-levels of automated data analysis to allow the engineers and human controllers to make crucial and well-informed decisions within seconds.
The second pivotal area in which big data are shaping space applications is their use for scientific purposes. As payloads are becoming more advanced, scientists find themselves overwhelmed with massive data sets, requiring advanced tools for collection, data processing, analysis, visualization and archiving (Hey; 2015), giving rise to what has been called a “data-intensive science”. The current level of analysis employed to process these SBD for science is cumbersome and time-consuming, often taking years before completion.
The increase in volume, variety and velocity of data during the general engineering design process calls for more advanced tools and methodologies, to increase efficiency, robustness and optimality, reduce lifecycle costs, and minimize errors (Wang & Alexander; 2015). Hence, in addition to telemetry tracking, mission operations and scientific purposes, SBD is becoming increasingly important during the design of space systems. As the space supply chain continues to grow and the industry moves more towards true commercialization, this need will become essential for enterprises to maintain their competitiveness and secure market share.
The added value of satsearch
The growing challenges faced by the space engineering community in handling complex design data underpins our work to curate, harmonize, and structure supply chain knowledge. This is the first step in ensuring that the industry moves towards more advanced design methodologies like Model-Based Systems Engineering (MBSE) to support the development of increasingly complex space missions.
Supply chain knowledge sits at the core of the design process and is currently highly scattered, incomplete and unstructured. At satsearch, we are focussing our efforts on developing a richer, structured, standardized format of representation for supply chain knowledge. Building a knowledgebase using this format will enable deep, complex querying to support optimization, sensitivity analysis, and risk mitigation during the design process.
By developing a supply chain knowledgebase that utilizes our structured data format, we are enabling engineers to sort, filter, and compare thousands of space products at the parameter level. Such a powerful comparison platform can be easily integrated into cutting-edge high-tech tools for early-stage design phases, like RHEA Group’s CDP4™ and Valispace.
As touched upon in a previous blog article, frustration in manually searching through PDF datasheets and general lack of supply chain transparency makes it extremely difficult for engineers to evaluate the true performance of space products. In a slow and cumbersome document-centric approach, the engineer is left with the only choice of copy-pasting the numbers provided in the datasheet.
By fully capturing all the attributes of a datasheet in a model-based manner, the structured format underpinning the satsearch database allows engineers to evaluate space products in a much more effective manner. We use this structured format to generate Electronic Data Sheets (EDS) for products; effectively a better version of the traditional PDF datasheet for engineering. Our vision is to capture all supply chain knowledge embedded in interface documents, technical manuals, and more to take the pain out of the engineering, and put the smile back on engineers’ faces.
The true beauty of migrating supply chain knowledge to a rich, structured format is that it renders every attribute of a space product uniquely addressable. Designation of unique IDs for each attribute belonging to a product allows for advanced analysis to pinpoint products that satisfy complex and uncertain design requirements and constraints.
By adding reasoning and constraints to attributes, the analysis process is thus less error-prone, resulting in a more robust and stable comparison of products, down to individual parameters. Data analysis powered by satsearch enables enhanced data querying, leading to useful insights from heterogeneous data for engineering. This scales to SBD for design because the EDS structure enables algorithmic searching, filtering, sorting, and comparing.
The bottom line is that the structured format, i.e., EDS, used to store supply chain knowledge in our knowledge-base, allows us to capture all the parameters of a space product in a much richer manner. With a view towards enabling advanced mission design methodologies like AI-based decision support systems, the satsearch platform will see major developments in the coming years. Satsearch is committed to the goal of supporting pioneers to keep pushing the boundaries of space exploration by radically improving the way complex missions are designed.