Undergraduate research · 2023

UAV swarm energy optimization

Simulation-based investigation of UAV-node energy consumption using metaheuristics and machine learning, compared with Harris Hawks Optimization.

  • UAV research
  • AI training

The context

Battery constraints limit the useful lifetime of nodes in a UAV swarm. This undergraduate study investigated methods to predict and reduce node energy consumption, using Harris Hawks Optimization (HHO) as a comparison baseline. The work was supervised by Dr. Tamoor Shafique.

My contribution

  • Investigated energy-aware design for UAV swarm networks through simulation.
  • Designed and evaluated genetic algorithms, fuzzy logic, decision trees, and convolutional neural networks for energy-consumption prediction and optimization.
  • Implemented and compared simulations in MATLAB, Octave, and Python, analyzing candidate approaches against an HHO baseline.

Challenges & solutions

Investigating battery-life constraints across UAV swarm nodes.

Framed node energy consumption as a simulation-based prediction and optimization problem and explored both metaheuristic and machine learning methods.

Comparing candidate methods against an established approach.

Used Harris Hawks Optimization as a baseline and implemented comparative simulations in MATLAB, Octave, and Python.

Pipeline overview

A simplified view of the workflow.

  1. Swarm energy scenario
  2. Optimization & ML methods
  3. Comparative simulation
  4. HHO baseline analysis

Scope & outcomes

Undergraduate research
2023
Evaluation environment
Simulation
Comparison baseline
HHO

Completed simulation-based undergraduate research, as documented in the research CV. No flight deployment, publication, or numerical improvement over the baseline is claimed in this overview.

Building something related?

Tell me about the problem, the constraints, and what success would look like.

Discuss a similar project

Explore more work

Local AI runtime & agent orchestration

Localbots

A Go-based local agent framework and web studio: a manager coordinates specialist bots through one resident Qwen3 model, with no cloud API required after setup.

Computer vision & data engineering

Skin-tone & dataset pipelines

Skin-tone and undertone analysis, a custom Monk Skin Tone detection system, and model-ready image data workflows.