ByBRAIN — Ayam JagoAutonomous Thermal Inspection

Self-operating
thermal inspection.

Every day, on the same route.

RoboGrid Inspect is an autonomous inspection robot designed for panel rooms, indoor substations, and battery rooms. It navigates pre-mapped routes, reads each panel using thermal and RGB cameras, and streams real-time asset condition evidence straight to the dashboard.

RoboGrid Inspect Robot
Thermal Heatmap · LV PanelLive
Max 72.4 °C

Built by

BRAIN — Ayam Jago

Indonesian AIoT & robotics

Targeted Operating Environments

Panel RoomIndoor SubstationBattery RoomUPS RoomData CenterIndustrial FacilityEV Charging DepotCommercial BuildingPanel RoomIndoor SubstationBattery RoomUPS RoomData CenterIndustrial FacilityEV Charging DepotCommercial Building

≥95%

Mission success rate

in test environment

≤30 cm

Waypoint positioning error

stop accuracy

≥85%

Detection precision / recall

defined scenarios

<1 s

Alarm latency

on local network

≥30%

Patrol time reduction

compared to manual

Prototype targets and pilot OKRs follow the Definition of Done in the product specification.

Problems

Manual patrols are costly, repetitive, and difficult to verify.

Repetitive manual patrols

Personnel walk the same routes in panel rooms and energized areas to read indicators and check for thermal hotspots.

Blind intervals between inspections

Anomalies that emerge right after an inspection are only caught on the next round — or after a failure occurs.

Inconsistent comparison data

Photos, thermal images, and technician notes are captured from inconsistent angles and times, making trends hard to trace.

Personnel constraints

High-value engineering staff spend hours on routine patrols that could be largely automated.

Consistent thermal evidence

Every panel is captured from the exact same pose on every visit, allowing temperature rise to be compared over time rather than guessed.

Reduced personnel exposure

The robot conducts initial sweeps in panel rooms and utility spaces, lowering engineer entry frequency in hazardous areas.

Edge-based analytics

Emissivity correction, hotspot detection, and anomaly scoring execute directly on-device for real-time alerting even on constrained networks.

How It Works

Six steps, from mapping to reporting.

01

Area mapping

Routes and asset waypoints are recorded once, then reused for every mission.

02

Scheduled missions

The robot performs automated patrols with obstacle avoidance and emergency stop features.

03

Panel validation

The gripper camera reads ArUco markers to confirm the correct panel before inspecting.

04

Data acquisition

The mast thermal camera and RGB camera record asset conditions from consistent poses every visit.

05

Edge processing

Emissivity correction, hotspot detection, and anomaly scoring run directly on the device.

06

Evidence & action

Findings appear on the dashboard with thermal proof, trend graphs, and exportable reports.

What the Robot Sees

Three perspectives, one mission.

RGB Camera — gripper

Reads ArUco markers, opens panel covers, and captures close-up inspection footage.

Thermal Camera — mast

Scans the entire panel surface and flags hotspots elevated above baseline.

Facility Navigation

Traverses pre-mapped routes, stopping precisely at every asset waypoint.

RoboGrid Inspect Hardware

Hardware

Built from accessible components.

Designed under a prototype budget of Rp7 million (~$450), utilizing a 4WD chassis, an edge computer, and a low-cost thermal array — not an industrial-grade robotics platform.

Thermal camera
MLX90640, mounted atop the mast
RGB camera
Mounted on the arm gripper
Navigation
ArUco markers + 9-axis IMU + wheel odometry
Obstacle detection
Short-range infrared sensors
Drive system
4WD chassis with motor encoders
Communication
Wi-Fi, MQTT, REST API

Dashboard

Actionable inspection evidence.

Mission Overview
Live

Missions

3/8

Panels

24/120

Avg

48.6°

Panel ID 23CRITICAL

Mission Overview

Thermal Heatmap
Rec
Thermal heatmap snapshot72.4 °C

Max

72.4°

Avg

46.2°

Base

42.0°

Thermal Inspection

Navigation Map
Auto
Navigation map snapshot
Battery78%
Mission progress62%

Navigation & Missions

Mission dashboard

Mission overview, live progress, and key KPIs consolidated on a single screen.

Thermal heatmap

Thermal mapping and hotspot detection per panel, complete with temperature scale.

Temperature trends

Asset thermal history so gradual temperature rises are identified before causing downtime.

Multi-level alarms

Warning and critical alerts with operator acknowledgment workflows.

Automated reports

Thermal inspection reports ready for PDF and Excel export for documentation.

AI Assistant

Query asset health, risky panels, and action recommendations directly on the dashboard.

Frequently Asked Questions

No. RoboGrid automates routine patrols and condition documentation. Findings are always verified by technicians, and those verification results serve as training data for future iterations.

No. RoboGrid is not a replacement for protection relays, SCADA, or fire/gas systems, nor does it diagnose internal faults from a single sensor alone.

Designed for indoor or semi-indoor environments with repeatable mappable routes: panel rooms, indoor substations, battery rooms, and industrial facilities.

Prototypes are tested on mockups or safe areas without exposed high voltage. Operation in energized environments requires HSE assessments, EMC compliance, and site access clearance.

Robot-as-a-Service with installation and subscription per site, direct hardware sales, SaaS analytics for history and reporting, and integration services for CMMS and asset management systems.

Not yet. This demo dashboard operates on simulated data to demonstrate the end-to-end workflow before physical hardware is installed on site.

See the workflow in action.

The demo dashboard runs on simulated live data — missions, thermal findings, alarms, and exportable reports, exactly as displayed when deployed on site.

Open Demo Dashboard