LoadZone, a curbside loading bay sensor concept

A low, battery-powered puck bonded to the road in each loading bay, using a magnetometer and LoRaWAN to report whether the bay is free. Designed as an open alternative to vendor curb data.

Build sheet

Status: Proof of concept, TRL 3
Sensing: Magnetometer
Radio: LoRaWAN
Licence: CERN-OHL-S-2.0 / MIT

Can I build this?Expert · AI 2/5 · $100–150

Difficulty: Expert
AI-buildability: 2/5
Time: A few months
Cost: $100–150

Parts

LoRaWAN radio module 3-axis magnetometer solar e-paper sign

Tools

3D printercasting mold materialssoldering ironepoxy potting compound

Starter prompt

Based on "LoadZone, a curbside loading bay sensor concept" on Afterware: https://afterware.xyz/p/loadzone-a-curbside-loading-bay-sensor-concept

Write firmware for an ESP32-based or LoRaWAN-enabled battery node that reads a 3-axis magnetometer, such as a QMC5883L, and detects a large steel object parked above it. Use a baseline calibration and a hysteresis threshold to report "free" or "occupied" as a LoRa uplink every 18 seconds, with a heartbeat for battery monitoring. Include a simple Python server that receives the uplinks and keeps the latest state per bay.

Estimates based on the original write-up. Building with an AI assistant? Connect it to Afterware (free MCP server).

Story

LoadZone is a proof-of-concept design for a sensor that sits in the middle of each vehicle slot in a curbside loading bay. The puck is a cast dome over potted off-the-shelf parts, bonded to the road like a raised pavement marker. A magnetometer detects the steel body of a parked vehicle, and a LoRaWAN radio reports "free" or "occupied" to a server, which publishes live bay state to drivers and occupancy and dwell times to the city. A solar e-paper sign on an existing pole is an option for showing status on the street.

The project is documented as a concept at TRL 3, with renders, a general arrangement drawing, sizing calculations and a build plan in the repository. The stated design goals are open firmware and data, a sensor that cannot capture faces or number plates, and a cell budget derated 40 % against a five-year target. The author says it was shaped by lessons from San Francisco's SFpark magnetometer pucks.

The repository lists the status as a proof of concept on paper, so no working prototype is described in the source.

Resources · 1

Repository