Frequency & order analysis
Break vibration and noise down by rotating speed and isolate the components that come from defects.
Algorithm Development
Have data to analyze or a problem to solve?
NeuroSense AI will recommend the right solution for you.
We develop algorithms that interpret signals from equipment and sites, such as temperature, humidity, vibration and noise. From building the data collection setup to lightweight AI that runs on sensors and devices, we deliver solutions that can actually be operated in the field.
Featured Algorithm Projects
Company PAnomaly detection 
Company HNoise diagnosis 
Company KCondition-based diagnosis
We've Solved Field Problems in Many Ways
Every machine, process and product is different, so we choose and combine how the signal should be read.
Time-frequency analysis
Unfold signals that change over time, such as acceleration and gear shifts, into spectrograms to find abnormal sections.
Anomaly detection from normal data
Learn only normal patterns, even without defect data, and catch signals that deviate from them.
Deep-learning classification & multi-read
Judge many short segments and combine them, so a single misread doesn't sway the result.
Noise & operating-condition compensation
Compensate for road, speed, temperature and mounting angle so the field is judged by the same standard.
Transfer learning & model compression
Reuse existing models to expand to new targets with little data, and slim them down to run on sensors and devices.
Algorithm Development Process
We work with you from collecting field data to integrating and operating the algorithm in your system.
- 01
Data collection setup
Collect field data from sensors, PLCs and vehicle networks.
- 02
Feature design
Find what separates states with frequency, order and statistical features.
- 03
Model development
Build and compare machine-learning and deep-learning models.
- 04
Mock-up & field validation
Validate on mock-up rigs and real field data.
- 05
Edge deployment
Slim the model down and deploy it on sensors and devices.
- 06
System integration
Connect to alarms, dashboards and existing systems.
Edge AI Algorithm Technology
We develop lightweight signal-processing and machine-learning algorithms that run directly on sensors and devices, judging on the sensor before sending data to cut traffic and power.
- 01Data
- 02Gathering & labeling
- 03Machine learning
- 04Intelligence pack
- 05Firmware build
- 06HW flash
Results produced on the sensor
Motion & state recognition
Tell operating states such as running, stopped, moving or stored from vibration patterns
Sound & vibration diagnosis
Judge defects by analyzing resonance, abnormal and rotating-part noise by frequency and order
Anomaly detection
Learn normal patterns, catch unusual signals and grade their risk
Event detection
Catch momentary events such as impacts, arcs and leaks and alert right away
Talk to us before you adopt.
- Review your data and requirements
- Field data collection and diagnosis if needed
- Algorithm development plan based on the review
Case Studies
We have technical know-how from applying algorithms across diverse industrial fields.
Company H
Automotive powertrain noise & vibration diagnosis
Company H
Real-time sync anomaly detection during durability tests
Company K
Fuel silo clog detection linked to automatic clearing
Company J
Inspection process quality analysis & AI reports
Company H
Industrial arc detection & risk grading
Company T
On-device AI for floor dirt detection in cleaners
Equipment Monitoring
- Steel-mill conveyor vibration monitoring
- Steel-mill temperature and humidity monitoring
- Development and delivery of pulley monitoring sensors
- Excavator monitoring system PoC
Sensors & Anomaly Detection
- In-house sensor integration with anomaly-detection algorithms
- Sensor data collection and system integration services
- Delivery of sewing-machine monitoring sensors
Energy & Plant
- Fuel-silo clogging detection solution for a power plant
- In-house IoT sensor for gas-pipeline impact and leak detection
Defense, Weapons Systems & Information Security
- CBM+ intelligent sensors for weapons systems
- Weapons-system monitoring and condition diagnostics
- Software anti-tampering technology and testbed
Smart Factory & DX/AX
- NeuroForge M.AX deployment in manufacturing processes
- DX·AX for inspection equipment on an automotive relay assembly line
- AI-based fastening-curve analysis for smart-factory quality prediction
- Non-intrusive PLC integration for real-time data collection in a paint shop
- Predictive failure solution for inline X-ray equipment
Mobility & Vehicle Diagnostics
- NVH sensors for vehicle-health diagnostic devices
- Vehicle fault-diagnosis sensors
- Turbocharger noise and fault diagnosis model
- EV powertrain anomaly-diagnosis PoC
R&D
- Smart electronic compass
- Battery-pack condition monitoring
- Autonomous-vehicle wheel safety sensor
- Vehicle-health smart sensor
- Predictive-maintenance IoT sensor
- Edge AI system-in-package (SiP)
- IoT sensor module
- NVH vibration and noise sensor
- Unsupervised-learning predictive maintenance
- IoT impact sensor for buried pipelines
- Booster-pump failure prediction solution
- Vision Zero transportation environment
- Support for hiring experienced R&D professionals