Research Projects

Cubesat formation flying for space debris characterization in LEO using aerodynamic forces

Background:
Characterizing debris and micrometeoroids of the sub-millimeter size in the LEO region is not possible by using ground based sensors. This demands in situ characterization, to make this characterization cost effective, a nanosat formation flying mission looks attractive. However, because of size, weight, and power (SWAP) constraints, the possibility of using perturbation forces without using an active thruster system for position control alleviates the constraints on SWAP.

Problem Statement:
To design a nanosat formation that enables characterization of sub-millimeter size objects in LEO using only the aerodynamic forces.

CHALLENGES POSED BY AERODYNAMIC FORCE BASED CONTROL:

- Limited control authority directly influences the position accuracy we can achieve.
- Uncertainty on the density leads to less robust control actions.
- Translation and attitude motions are coupled inherently.

Solution Approach:
- Formation flying will be designed by considering payload requirements.
- We expect to produce a reference trajectory that considers aerodynamic forces, optimized with respect to decay time that satisfies the mission position and control requirements and collision avoidance constraints.
- Followed by an inner control loop design that includes an estimator that estimates density profile.

Adaptable Architecture for Anomaly Detection in Communication Satellites using Machine Learning

Background:
In a distributed space system with multiple satellites, addressing anomalies or faults in a timely manner is crucial. Currently, the approach involves human intervention from ground control, leading to latency in response due to telemetry analysis and simulated fault conditions. To enhance efficiency, an adaptable system architecture can be developed.

Problem Statement:
Can a more adaptive system architecture be developed for distributed space systems(communication satellites constellation), allowing faster and more automated responses to anomalies?

Solution Approach:
To address the latency and improve fault response, an adaptable system architecture leveraging machine learning and digital twin solutions can be explored. The goal is to learn from past scenarios and enable autonomous decision-making. When a satellite experiences an anomaly, the system could benefit from insights derived from similar situations encountered by other satellites in the constellation utilizing Machine Learning algorithms.