On a quiet farm in rural Sweden, a fox once took an entire flock of chickens and ducks while the owners were away buying groceries. Four years later, a neighbor’s dog slipped through and did the same. These losses weren’t just financial—they were deeply personal. That’s why a tech-minded farmer turned to innovation, launching FenceGuard, an early alert system designed not to replace vigilance, but to enhance it.
The long-term vision is ambitious: a distributed network of LoRa sensors, edge AI running on TensorFlow Lite, real-time integration with Sweden’s weather service to filter out false alarms from wind and rain, and two product tiers—one for smaller predators like foxes and hawks, another for larger threats like wolves and bears. But vision alone doesn’t build systems. Good engineering demands validation before scaling, and that means starting where the unknowns are greatest: the data itself.
From problem to prototype: starting lean with what matters
The first step wasn’t a new purchase—it was reuse. A LoPy4 development board, already on hand from a previous weather station project, sits on a desk in a farm office. It’s not the latest or most powerful device, but it’s more than capable for early testing. The missing piece? A motion sensor. That’s exactly what arrived this week: a simple PIR (passive infrared) motion detector.
Using existing hardware isn’t laziness—it’s discipline. The goal isn’t to build fast; it’s to learn fast.
No new microcontroller. No full stack. Just one sensor and one board, connected, capturing data. The plan is to log every signal, every false positive, every gust of wind that shakes a branch. Before investing in robust infrastructure, the priority is understanding the raw signal—because sensor data is full of noise, and noise hides truth.
The real battle isn’t predators—it’s environmental chaos
A PIR sensor doesn’t distinguish a fox from a swaying branch. It detects temperature changes in its field of view. That’s all. And in an outdoor setting, that’s not enough. Wind moves trees. Rain bends grass. Temperature shifts alter sensitivity. A dog trotting past may trigger the same response as a human walking slowly.
The first question isn’t whether the system will work—it’s whether the sensor can even tell the difference between danger and the environment.
Over the next few weeks, the focus is narrow and unsentimental: collect data. Not predictions. Not alerts. Just raw motion events paired with timestamps, temperature readings, and environmental observations. The goal is simple: can a single PIR sensor, with no filters or AI, reliably separate real threats from daily noise?
The answer will shape everything that follows. If the data is clean, the path is clear. If it’s noisy, the real work begins: designing timing windows, threshold filters, and maybe even pairing motion with wind and temperature sensors to reduce false alarms. But none of that can be decided in theory. It must be decided in practice.
What comes next—and how to join the journey
Once the PIR sensor is wired to the LoPy4 and the first logs begin streaming, the next phase starts: publishing the setup, sharing the raw data, and inviting scrutiny. No polished app. No cloud dashboard. Just transparent results—because if FenceGuard is to be trusted, trust must be earned in the open.
This isn’t just a tech project. It’s a development diary. Hardware choices, firmware decisions, data pipelines—all documented in real time. The hope is to spark conversation among others building similar systems: livestock monitoring, perimeter security, outdoor IoT. If you’ve grappled with sensor noise, filtering logic, or edge deployment, your insights are valuable.
The path forward is iterative. Small steps. Clear questions. No shortcuts. And when the system finally sounds an alert that saves a life—whether chicken, duck, or trust in technology—it will do so not because the tech was perfect, but because the human behind it understood the noise before chasing the signal.
AI summary
İsveçli bir geliştirici, tavuk ve ördeklerini yırtıcı hayvanlardan korumak için PIR sensörüyle başlayan FenceGuard projesini nasıl hayata geçiriyor? Detaylar burada.
