Case Studies

Real-world robotics, automation and AI engineering projects delivered by Estalvo across Europe. Each case focuses on measurable impact, reliability in the field and long-term maintainability.

1. Autonomous Inspection with Unitree Go2

Industry: Renewable Energy · Location: Valencia, Spain

Challenge

A large solar farm needed regular panel inspection for hot-spots, damaged strings and foreign objects. Manual inspection was slow, expensive and only partially documented.

Solution

Unitree Go2 with thermal camera and 3D LiDAR payload.
High-resolution SLAM map of a 12-hectare solar field.
AI-based thermal anomaly detection pipeline with event logging.
Automated patrol missions configured in Estalvo Fleet Management Console.

Results

≈68% faster complete inspection cycles compared to manual procedures.
24/7 autonomous patrol capability with scheduled and on-demand missions.
Early detection of overheating panels and connectors, reducing downtime.
Return on investment (ROI) achieved in approximately 11 weeks.

2. Heavy-Duty Quadruped Logistics (Unitree B1)

Industry: Automotive Components · Location: Zaragoza, Spain

Challenge

Components weighing between 15 and 35 kg had to be transported between warehouse zones through narrow corridors, ramps and uneven floors, while maintaining strict takt times.

Solution

Unitree B1 configured with adaptive gait and terrain profiling.
3D obstacle avoidance with LiDAR + depth fusion.
Integration with the client's WMS/ERP via secure API for task dispatching.
Fleet-level traffic coordination to avoid congestion in bottleneck areas.

Results

≈42% reduction in internal logistics labour requirements.
Measured reduction in near-miss incidents and manual handling risks.
Stable operation on uneven industrial floors and loading ramps.
Delivery accuracy above 99.2% over the first three months of operation.

3. Energy-Efficiency AI Module for Robotic Cells

Industry: Industrial Robotics · Location: Bilbao, Spain

Challenge

Multiple robotic cells were configured to run at maximum power regardless of payload or cycle. Energy consumption and thermal stress were higher than necessary.

Solution

AI models trained on torque, RPM and duty-cycle telemetry per axis.
Dynamic power and speed profiles based on predicted load and takt time.
Energy-aware trajectory planning integrated into existing robot programs.
Thermal forecasting to drive proactive cooling and maintenance planning.

Results

Measured 18–24% reduction in energy use across selected cells.
Estimated 16% increase in component lifetime due to lower thermal stress.
More stable cycle times and improved takt adherence.
CO₂ savings exposed via management dashboards for sustainability reporting.