Assessments to Prepare and De-Risk Technology
Description: Assessments to Prepare and De-Risk Technology Developments Intelligent Space Camera Contractor(s): Ubotica (IE) GSTP Element 1 De-risk ESA Budget: 145 k Background and justification: The recent advent of on-board AI implemented with COTS
Related Topics
Download Presentation
"Assessments to Prepare and De-Risk Technology" is the property of its rightful owner. Permission is granted to download and print the materials on this website for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Presentation Transcript
slide1. Assessments to Prepare and De-Risk Technology Developments – Intelligent Space Camera Contractor(s): Ubotica (IE) GSTP Element 1 De-risk ESA Budget: 145 k€ Background and justification:
The recent advent of on-board AI implemented with COTS devices has enabled a push of intelligence to the satellite, processing data at source, but this has not yet extended to the domains of Visual Monitoring Cameras (VMCs)
Previous developments (e.g., as demonstrated in the PhiSat missions) have shown that low-power, high-performance AI compute is achievable and applicable using the Myriad family of Vision Processing units, which have previously undergone radiation characterization in collaboration with ESA
Current VMC and launch vehicle camera systems tend to be ‘data gatherers’ rather than intelligence gatherers – where the former is a process of visual data acquisition from a camera, followed by storage and/or transmission. This leads to the acquisition of significant amounts of 2D data that may ultimately only be of sporadic value
This activity addresses the lack of intelligence in current VMC and launch cameras by adding efficient AI compute, along with hardware-accelerated stream processing, as part of a new intelligent camera designed for space Objective(s):
To develop an Intelligent Space Camera that integrates state-of-the-art image quality with real-time Artificial Intelligence (AI) analysis of a video stream, using COTS technology
To build and test the Intelligent Space Camera hardware and software via a full prototype
To demonstrate an example use case using the Intelligent Space Camera
To verify the feasibility of the Myriad X for use in a compact form factor design in order to de-risk a follow-on full camera development activity Achievements and status:
Hardware design of compact form-factor camera completed
Multi-stack camera hardware solution manufactured and functional and electrical bringup completed
Software developed to enable RTSP streaming, ISP, H.265 compression, text overlay, Ethernet C&C, and user configuration via CURL
Demonstration application developed to execute on the hardware
Demonstration of streaming AI ongoing Benefits:
The added value of this technology for the targeted application (ESA mission, commercial application, etc.).
Small form-factor camera solution, operating in low power (~<10W) envelope and with Ethernet connectivity, can be easily deployed on many missions from CubeSats to larger satellites and launchers
The H.265 hardware accelerated compression, along with fully software programmable on-board processing, provides a flexible solution that can be adapted to VMC and launch camera applications
Command and control over Ethernet, facilitated via an on-camera HTTP server, enables control and data/intelligence acquisition from the camera without requiring any custom host software Next steps:
Completion of AI streaming optional functionality
Development of flight-ready hardware and software for VMC, targeting ESA/commercial missions from 2025, funded under a proposed GSTP follow-on activity
Initial IOD flight in 2024, in preparation for full flight qualification by 2025 Target TRL: 8 Date: Q2 2025 TRL Achieved: 4 YoC: 2023 Initial: 2 TO: Gianluca Furano (TEC-EDD) 4000134612/21/NL/GLC/va Intelligent Space Camera main stack Sample camera output showing overlay<br>
The recent advent of on-board AI implemented with COTS devices has enabled a push of intelligence to the satellite, processing data at source, but this has not yet extended to the domains of Visual Monitoring Cameras (VMCs)
Previous developments (e.g., as demonstrated in the PhiSat missions) have shown that low-power, high-performance AI compute is achievable and applicable using the Myriad family of Vision Processing units, which have previously undergone radiation characterization in collaboration with ESA
Current VMC and launch vehicle camera systems tend to be ‘data gatherers’ rather than intelligence gatherers – where the former is a process of visual data acquisition from a camera, followed by storage and/or transmission. This leads to the acquisition of significant amounts of 2D data that may ultimately only be of sporadic value
This activity addresses the lack of intelligence in current VMC and launch cameras by adding efficient AI compute, along with hardware-accelerated stream processing, as part of a new intelligent camera designed for space Objective(s):
To develop an Intelligent Space Camera that integrates state-of-the-art image quality with real-time Artificial Intelligence (AI) analysis of a video stream, using COTS technology
To build and test the Intelligent Space Camera hardware and software via a full prototype
To demonstrate an example use case using the Intelligent Space Camera
To verify the feasibility of the Myriad X for use in a compact form factor design in order to de-risk a follow-on full camera development activity Achievements and status:
Hardware design of compact form-factor camera completed
Multi-stack camera hardware solution manufactured and functional and electrical bringup completed
Software developed to enable RTSP streaming, ISP, H.265 compression, text overlay, Ethernet C&C, and user configuration via CURL
Demonstration application developed to execute on the hardware
Demonstration of streaming AI ongoing Benefits:
The added value of this technology for the targeted application (ESA mission, commercial application, etc.).
Small form-factor camera solution, operating in low power (~<10W) envelope and with Ethernet connectivity, can be easily deployed on many missions from CubeSats to larger satellites and launchers
The H.265 hardware accelerated compression, along with fully software programmable on-board processing, provides a flexible solution that can be adapted to VMC and launch camera applications
Command and control over Ethernet, facilitated via an on-camera HTTP server, enables control and data/intelligence acquisition from the camera without requiring any custom host software Next steps:
Completion of AI streaming optional functionality
Development of flight-ready hardware and software for VMC, targeting ESA/commercial missions from 2025, funded under a proposed GSTP follow-on activity
Initial IOD flight in 2024, in preparation for full flight qualification by 2025 Target TRL: 8 Date: Q2 2025 TRL Achieved: 4 YoC: 2023 Initial: 2 TO: Gianluca Furano (TEC-EDD) 4000134612/21/NL/GLC/va Intelligent Space Camera main stack Sample camera output showing overlay<br>