Title: Smart Camera System On-board a CubeSat for Space-based Object Reentry and Tracking
Author(s): Ravi Teja Nallapu, Aaditya Ravindran, Himangshu Kalita, Vishnu Reddy, Roberto Furfaro, Erik Asphaug, Jekan Thangavelautham
Published in: Proceedings of IEEE/ION PLANS 2018
April 23 - 26, 2018
Hyatt Regency Hotel
Monterey, CA
Pages: 1294 - 1301
Cite this article: Nallapu, Ravi Teja, Ravindran, Aaditya, Kalita, Himangshu, Reddy, Vishnu, Furfaro, Roberto, Asphaug, Erik, Thangavelautham, Jekan, "Smart Camera System On-board a CubeSat for Space-based Object Reentry and Tracking," Proceedings of IEEE/ION PLANS 2018, Monterey, CA, April 2018, pp. 1294-1301.
Full Paper: ION Members/Non-Members: 1 Download Credit
Sign In
Abstract: The availability of low-cost, high-performance electronics, sensors, actuators, and communication systems for CubeSats and small-satellites open new avenues in Space Situational Awareness (SSA). Scores of low-cost CubeSats and small-satellites maybe rapidly assembled into a constellation to constantly observe natural and man-made objects entering the Earth’s atmosphere. Such a low-cost architecture can be constantly upgraded with rapidly advancing sensor technology and is robust to single point failures. SWIMSat (Space Weather and meteor Impact Monitoring Satellite) is a Low Earth orbit (LEO) based 3U CubeSat demonstrator concept intended to observe objects entering the Earth’s atmosphere. The proposed spacecraft uses a smart camera system to observe the Earth during the night time. The camera is placed on the long axis of the spacecraft so that the earth observation is achieved with Nadir pointing. Using this demonstrator, we hope to develop the technology to demonstrate a CubeSat constellation to observe object entering the Earth’s atmosphere. SWIMSat’s smart camera system uses an onboard image processing algorithm to detect objects entering the Earth’s atmosphere. The object detection algorithm uses a multilayered approach to autonomously detect and track entering objects using a single camera. The object detection approach at its base relies on simple vision-feature detection methods to filter and identify events of interest. This is followed by obtaining dynamic information of the objects including velocity, acceleration, trajectory. Using the hyperspectral and thermal imagers it may be possible to obtain first-order estimates of the composition of the reentering object. Using these estimated statistics, our approach develops a physical simulation model of the observed system and predicts its entry trajectory. In this work, the performance of our object entry detection algorithm coupled with a spacecraft guidance, navigation, and control system is demonstrated by simulating both the physical world and the orbiting observer. These experiments are conducted in the laboratory using a mock spacecraft mounted on a robotic arm to facilitate 2-axis rotations. These two systems are connected to a computer equivalent to a CubeSat computer. When the experiment starts, the camera detects incoming meteors simulated on a high resolution TV-screen. The detection algorithm coupled with spacecraft control system then track and predict the object reentry trajectory. A thorough description of the detection algorithm, along with the tracking controller is presented in this work. The results of laboratory hardware-in-the-loop experiments are presented. Our work suggests both a critical need and the promise of such a tracking algorithm for implementation of an autonomous, low-cost constellation for performing Space Situational Awareness (SSA).