Inspiration
Modern devices constantly move between different connectivity options such as Wi-Fi and 5G. A simple strongest-signal approach can cause frequent switching when two networks have similar quality, leading to unstable connections and unnecessary handovers.
This inspired SeamWise – EDGE, a network-selection concept designed to make switching decisions more intelligently by considering several network conditions instead of signal strength alone.
What it does
SeamWise compares available networks using multiple metrics, including:
- Signal quality
- Latency
- Estimated packet loss
- Hysteresis to prevent unnecessary switching
Instead of immediately changing networks whenever another signal becomes slightly stronger, the system waits until the alternative provides a meaningful improvement. This helps reduce frequent back-and-forth switching, often called the ping-pong effect.
How we built it
We created a Python-based simulation using NumPy and Matplotlib and tested it in Jupyter Notebook.
The simulation models changing Wi-Fi and 5G conditions over time. We compared three approaches:
- Signal-strength-only network selection
- Signal selection with hysteresis
- Our proposed multi-metric selection algorithm
The proposed algorithm calculates a score based on signal quality, latency and estimated packet loss, while also applying a hysteresis margin before allowing a handover.
We first tested a single scenario to visualize how the networks changed over time. We then expanded the evaluation to 30 different simulated scenarios using different random seeds to check whether the behavior was consistent.
We measured metrics such as:
- Number of network switches
- Average latency
- Estimated packet loss
- Modeled cumulative handover interruption
For the interruption model, we used an assumed fixed delay per handover, so these values represent simulated estimates rather than real-world measurements.
What we learned
One of the main things we learned was that choosing the strongest signal is not always the best decision.
A network may have a slightly stronger signal while still having worse latency or stability. We also learned that reducing unnecessary handovers can improve connection stability even when other metrics remain similar.
Another important lesson was the importance of comparing against a fair baseline. Instead of comparing our solution only with a basic strongest-signal method, we also introduced a stronger signal-plus-hysteresis baseline.
Challenges we faced
The biggest challenge was creating a simulation that represents changing network conditions without access to a full real-world Wi-Fi/5G test environment.
We also had to balance different metrics. Increasing stability too much can cause the device to remain on one network longer than necessary, while switching too aggressively can create repeated handovers.
Another challenge was distinguishing between simulated estimates and measured network performance. We therefore clearly separate modeled handover interruption and estimated packet loss from real-world measurements.
What's next
The next stage would be to test SeamWise using real network interfaces and live traffic measurements. We would also like to evaluate additional factors such as bandwidth, jitter, congestion and application requirements.
Ultimately, SeamWise could evolve into an adaptive connectivity system that makes smarter network decisions while users move between different communication technologies.
