Closed Loop Artificial Intelligence Cognitive Synthetic Aperture Radar

• Country: Netherlands
• Summary: Closed Loop Artificial Intelligence Cognitive Synthetic Aperture Radar
• NLT Ref No: 98806569
• Deadline: 08 May 2024
• Financier: Other Funding Agencies
• Purchaser Ownership: Government
• Tender Value: Refer Document
• Notice Type: Tender
• Document Ref. No.: 1-12142
• Purchaser's Detail:
  Login to see full purchaser details. Login to see full purchaser details. Login to see full purchaser details. Login to see full purchaser details. Login to see full purchaser details. Login to see full purchaser details. Login to see full purchaser details. Login to see full purchaser details. Login to see full purchaser details.

  Login to see details
• Description:
• Tenders are invited for Closed Loop Artificial Intelligence Cognitive Synthetic Aperture Radar Price Range 500 Keuro The availability of powerful on-board processing devices is enabling new possibility to develop cognitive microwave instrument and in particular radar and Synthetic Aperture Radar. Although a lot of work is currently underway on Machine Learning (ML) for processing sensor data, the potential use of ML in closed-loop cognitive radar is not fully exploited yet. A cognitive radar can adapt its operational mode based on the scenario/environment monitored, adjusting waveform parameters such as frequency, pulse width, PRI(pulse repetition interval) (PRI), transmitted power, up to the transmit and receive antenna patterns (in case the radar instrument makesuse of phased array antennas). The activity objective is to implement an on-board AI data analytics for rapid and resilient (near-real-time) crisis response, e.g. maritime safety and security, surveillance and reconnaissance, earthquakes, flooding, especially if data fusion with different payloads data (e.g. AIS) is possible. Spot applications are also possible for space debris detection and classification (e.g. size). The most interesting example of application could be the use of reduced RF bandwidth (BW) and high coverage SAR mode (e.g. SCANSAR) in order to simulate a "scanning" operation typical of ground military radars. Once specific target/feature has been identified by the AI, the radar instrument changes its operational mode in order to use the full available instrument resources (RF BW, RF Power) to acquire a small vignette (e.g. wave mode/reduced spotlight) where the identified target with high resolution and high SNR for better recognition is imaged. In this way, the full SAR payload capability will be used only in presence of the specific "targets", for which the AI has been trained. A careful application scenario analysis, evaluation and demonstration of ML methods is therefore in need. Evaluating ML methods for a set of reference scenarios will pave the way for their deployment in future Synthetic Aperture Radar instruments. Another application for cognitive radar is to enable an innovative method for on-board dynamic generation of the radar signals to adapt SAR-based payload instrument with respect to earth observation tasks and todata acquisition and operation of dynamic schedule scenarios. This capability will entail the implementation and exploitation of aset of key functions, such as: adaptive SAR sensing, dynamic waveform generation, real-time closed-loop radar, compressivesensing (sub-Nyquist sampling) autonomous tracking, efficient data compression techniques, on board SAR focusing algorithms.The activity shall define: suitable application scenarios enabled by on-board AI processing, on-board processing algorithms, SAR system closed-loop operational concepts, system and instrument architectures, required technological developments.The activity shall implement a representative hardware (HW) breadboard implementing selected algorithms to demonstrate the achievable on-board capabilities for a subset of representative application scenarios and shall encompasses the following tasks: Definition of scenarios, target and high-level application that can benefit from the usage of on-board AI analytics of SAR raw/preprocessed data. Selection, design and optimisation of ML model(s) for the defined scenarios, and generation/gathering of training datasets, taking into account reference scenario and pre-processing needs. Selection of SAR modes to be used in an adaptive framework, considering acquisition scenario and on-board (raw) data exploitationneeds through ML models. Definition of an agile radar instrument architecture and relative HW, whose functionalities and capabilities can fulfil the need of implementing different operational SAR mode withthe required low latency. Definition of possible technological development needs necessary for the implementation of the requiredinstrument adaptability/scalability functionalities and/or processing capabilities. Selection of representative ML HW accelerator(s), and inclusion in a breadboard for demonstration. Validation of trained model(s), with real/synthetic data, and evaluation ofthe results Deployment of trained model on accelerator HW and demonstration of validation results, as well as benchmarking.NOTE: This activity is currently in the GSTP Work Plan but will only be implemented after confirmation of financial support from Delegations. The final list of participating states will be known at tender issue.
• Documents:

 Tender Notice