Skip to content

Industrial Sensor + IoT

Adaptive sensor intelligence for the machines that run modern industry.

Adasens designs adaptive multi-modal sensor stacks — LiDAR, mmWave radar, ToF imaging, and edge-AI fusion — that cut perception latency by up to 68% and ship to production in 14 weeks. We work shoulder-to-shoulder with engineering teams from prototype to FAA, CE, and UL certified deployment.

  • Trusted by
  • John Deere
  • ABB
  • Continental
  • Oshkosh
  • + 10 other Fortune 500 manufacturers

Performance envelope

  • −68% Sensor-fusion latency 240+ deployed units, 2024
  • 14 wk Prototype to production vs. ~11 mo industry avg
  • <4 W Power per sensor node Edge AI Review, Q3 2024
  • ISO + AS 9001:2015 & AS9100D Boulder · Stuttgart

Sensor modalities

One fusion engine. Four sensing modalities.

Each modality is engineered in-house and fused at the edge. No off-the-shelf black boxes — every channel runs through our adaptive weighting model on under 4 watts per node.

LiDAR sensor emitting concentric range rings onto a measurement grid.

M-01

LiDAR

905 nm and 1550 nm time-of-flight heads with adaptive scan patterns. 0.05° angular resolution, 200 m range class, IP67 housings for cab-mounted or mast-mounted deployment.

Wavelength
905 / 1550 nm
Range
up to 200 m
Resolution
0.05° angular
Interface
Ethernet / CAN-FD
mmWave radar module emitting radio wave arcs from a small printed circuit board.

M-02

mmWave Radar

76–81 GHz FMCW front ends purpose-built for dust, rain, and thermal bloom. Sub-degree AoA with on-chip DSP, fused natively into our edge pipeline.

Band
76–81 GHz
Modulation
FMCW / PMCW
AoA accuracy
< 1°
Latency
8 ms typical
Time-of-flight depth camera producing a point cloud of an industrial part.

M-03

ToF Imaging

Solid-state ToF arrays from VGA to 1.2 MP, depth-compressed on-chip. Designed for bin picking, operator presence, and cabin monitoring where dense near-field perception matters.

Resolution
VGA → 1.2 MP
Depth
0.2 – 8 m
Frame rate
up to 60 fps
Output
depth + IR + RGB
Edge-AI fusion diagram showing LiDAR, mmWave, and ToF inputs converging into a single track and classify output.

M-04

Edge-AI Fusion

A small, deterministic fusion model — not a foundation model — running on under 4 W per node. Adaptive weighting learns the modality mix per environment without retraining the stack.

Power
< 4 W / node
Runtime
ONNX / TensorRT
Latency
−68% vs. baseline
Targets
Jetson · ARM · x86

By the numbers

Measured across 1,800+ commercial deployments since 2017.

  • 1,800+ Commercial systems deployed Across 42 countries
  • 97.4% On-time delivery rate Production-grade sensor kits
  • 2.4B Sensor events processed daily Active installed base
  • 87 Engineers on staff 41% with advanced degrees
  • 14 Fortune 500 customers John Deere · ABB · Continental · Oshkosh
  • 19 ROS 2 + AUTOSAR platforms No middleware rewrites required
  • 3.2× Faster iteration cycles In-house anechoic & EMC chambers
  • 2 Engineering offices Boulder, CO · Stuttgart, DE

Engagement model

From kickoff call to certified production in 14 weeks.

No black box. Every step has a written deliverable you can review with your hardware lead.

  1. Week 0

    Discovery call

    Deliverable: written environment model and 30-minute technical consultation summary.

    No NDA required for the first call. We sign mutual NDAs before any data is shared.

  2. Week 1–3

    Reference design

    Deliverable: modality selection, BOM, latency and power budget in your stack.

    Reviewed in a shared engineering channel with your EE and firmware leads.

  3. Week 4–8

    Bench integration

    Deliverable: sensor kit on your bench, fused output on your middleware, anomaly report.

    Runs against your existing ROS 2 or AUTOSAR environment without rewrites.

  4. Week 9–14

    Certified production ramp

    Deliverable: FAA / CE / UL paperwork, production-test fixture, on-time delivery plan.

    Ramped in our ISO 9001:2015 and AS9100D certified facilities.

Middleware compatibility

Plugs into the middleware you already run.

Integrated across 19 ROS 2 distributions and AUTOSAR platforms without middleware rewrites — verified on the bench, not on a slide.

ROS 2 distributions

  • ROS 2 Humble Hawksbill
  • ROS 2 Iron Irwini
  • ROS 2 Jazzy Jalisco
  • ROS 2 Kilted Kaiju
  • Micro-ROS (embedded)
  • Apex.OS

AUTOSAR platforms

  • AUTOSAR Classic (4.4 → R21-11)
  • AUTOSAR Adaptive (R21-11)
  • EB tresos
  • Vector DaVinci
  • ETAS RTA-CAR
  • Bosch RCS

Edge runtimes

  • NVIDIA Jetson (Orin · Xavier)
  • ARM Cortex-A / R-class SoCs
  • x86 industrial PCs
  • ONNX Runtime / TensorRT
  • Vendor SDKs (Intel / NXP / TI)
  • Custom Yocto layers

Vehicle + machine buses

  • CAN 2.0B / CAN-FD
  • 100BASE-T1 / 1000BASE-T1
  • RS-485 / RS-422
  • EtherCAT
  • LIN
  • PCIe / USB 3.2

// no middleware rewrites. no black boxes. drop-in sensor stack.

Certifications

Both ISO 9001:2015 and AS9100D in-house.

Only sensor-fusion vendor certified to both standards. Our Boulder and Stuttgart facilities share the same QMS, the same test fixtures, and the same engineering leads.

ISO 9001:2015 Quality management — all sites Current
AS9100D Aerospace-grade QMS Current
CE / UL / FCC Product safety + EMC Per SKU
FAA conformance UAS sensor payloads Per program

Awards & recognition

Recognized by the engineering press.

  • 2024 Edge AI Review — Top 10 Perception Vendors
  • 2024 Robotics Business Review — Sensor Partner of the Year
  • 2023 Robotics Business Review — Sensor Partner of the Year
  • 2022 Embedded Vision Alliance — Engineering Excellence

“We replaced three separate sensor vendors with one Adasens fusion stack. The latency number on the bench matched the latency number in the field — which is not something I’m used to writing in a design review.”

M. Halvorsen Principal Hardware Engineer, Industrial OEM customer

Talk to an engineer

Book a 30-minute Sensor Strategy Call.

Bring your modality mix, your latency budget, and your integration timeline. We'll come back with a reference design, a BOM, and a 14-week path to certified production.