All Projects
IoT2024Production

Sentinel

IoT Fleet Monitoring Platform

Real-time monitoring platform for logistics fleets across Europe. Processes 2M+ sensor events per day from 800+ connected vehicles, enabling predictive maintenance and proactive incident response.

800+

Vehicles Monitored

2M+

Daily Events

34%

Downtime Reduction

Context

A mid-size European logistics operator was flying blind across its 800-vehicle fleet — incident reports arrived hours after the fact, maintenance was reactive, and fuel cost overruns were untracked. Leadership needed operational visibility in real-time.

The Challenge

Vehicles traversed 12 countries with intermittent GSM coverage. Data came from 7 different sensor manufacturers with incompatible protocols, and the existing data warehouse was batch-only with 6-hour latency — far too slow for operational decisions.

The Solution

I designed and led the build of a unified edge-to-cloud IoT architecture. Each vehicle received a hardened gateway running a lightweight MQTT broker and custom firmware for protocol normalization. Events streamed to AWS IoT Core and landed in InfluxDB for real-time querying. A Grafana-based operations console gave dispatchers live dashboards per vehicle, per route, and per driver.

Methodology

  • Edge-first architecture to handle poor connectivity
  • Protocol abstraction layer to normalize heterogeneous sensor output
  • Time-series data model optimized for write-heavy telemetry workloads
  • Alert rules engine for threshold-based and anomaly-based incident detection

Impact

Deployed to full fleet within 4 months. Incident response time dropped from 6 hours to under 8 minutes. Predictive maintenance model (trained on the first 3 months of telemetry) enabled a 34% reduction in unplanned downtime.

Technology Stack

PythonMQTTInfluxDBGrafanaAWS IoT CoreKubernetesEdge Computing

Project Details

CategoryIoT
Year2024
StatusProduction