Friland - Monitoring and control system for off-grid tiny houses

Summary

An integrated telemetry, control and remote-management system for off-grid living modules (mobile tiny houses) operated by an Italian experiential-tourism startup. Custom electronics designed in-house, C/C++ firmware on ESP32, direct integration with inverter, MPPT and battery storage over Modbus TCP, 4G telemetry to a cloud IoT platform. Six years in production: 12 units still running on their original electronics, out of a fleet of 20 modules deployed across Friuli, Tuscany and Emilia.

Client

Friland (fri.land), a hospitality startup founded in 2019. The first cabin opened to the public on 12 July 2020. Mobile tiny houses spread across several Italian regions, placed in isolated natural settings and periodically relocated. Very short stays (1-3 nights on average), frequent turnover, different guests every cycle. The technical collaboration with the client has been ongoing since 2018.

Problem

The modules are designed to sit in remote locations with no connections at all: no power grid, no water main, no sewer. Each cabin produces and consumes its own energy (solar PV + storage), manages its own water (separate tanks for clean, grey and black water), heats, cools and handles guest access. Without a central telemetry and control system the operating model falls apart: a fault, a flat battery or a full tank discovered when the guest arrives means cancellation, refund and reputational damage; at the same time, sending a technician blind to modules spread across three regions is economically unsustainable. The system had to prevent failures before they became visible to guests, and turn maintenance from reactive into planned.

Solution

A three-layer architecture, designed and built in-house from the electronics to the cloud.

On-site electronics. A main PCB designed from scratch (schematic and layout) around the ESP32, with an integrated sensing chain: magnetic sensors on doors and windows, analog current sensors, ultrasonic probes for tank levels, one-wire temperature probes, flood sensors, flow meters, and relays for load control (heating, cooling, winter anti-freeze cables, access systems). Advanced telemetry pulled directly from the inverter over Modbus TCP: battery state of charge, PV production, MPPT status, inverter status, component temperatures, currents and voltages. A second PCB drives a paper display in the module’s living area, showing guests their water and energy consumption and the cabin’s solar production in real time - a transparency and awareness feature as much as an engagement one.

Connectivity. An industrial-grade 4G router runs a protected LAN inside the module; MQTT communication to the cloud. An offline-first architecture: the firmware logs telemetry locally (minute-level sampling on most values) and publishes it in bulk as soon as the network is available, with no data loss.

Cloud and operations. Ingest, storage, dashboards, alerting rules and notifications on the Things5 enterprise IoT platform. Friland staff receive real-time alerts on water levels (the most frequent alert category), battery state, abnormal temperatures and other relevant events, and plan interventions accordingly. A dedicated field tool written in Go records the work carried out and keeps a per-module history.

Results

• 20 tiny houses installed in total; 12 still operating on their original 2020 electronics.
• Six years of continuous operation in remote environments, through extreme seasonal thermal cycles.
• Variable placement: the cabins move, the infrastructure follows them with no manual reconfiguration.
• Minute-level telemetry stored in the cloud, real-time alerts to staff.
• Maintenance shifted from reactive to planned, cutting out unnecessary call-outs.

Why it worked

A system like this only holds up if whoever designs it commands the whole stack - from the circuit board to the firmware to the cloud - because every local decision has consequences for operation years later. Reading the inverter directly over Modbus TCP, instead of relying on the manufacturer’s cloud APIs, guaranteed fine-grained telemetry, vendor independence and continuous operation even when third-party services are unavailable. The offline-first architecture was designed from day one for the real operating context, where 4G coverage is intermittent. Six years in production with no work on the original electronics, in thermally and mechanically hostile environments, is the real measure of design quality.