We build the fusion that keeps your system navigating

Q: What is sensor fusion in positioning and navigation?

Sensor fusion combines measurements from multiple sensors - such as UWB ranging, an IMU, and LiDAR or visual odometry - into a single estimate of position, velocity, and orientation that is more accurate and more reliable than any sensor alone. The point is resilience: each sensor's failure mode is covered by another's strength. needCode designs and implements navigation-grade fusion for robots, vehicles, and tracked platforms, engineered to run on embedded hardware.

Q: Why fuse UWB with an IMU?

UWB and an IMU are natural complements: UWB provides absolute position but at a limited update rate and with occasional bad measurements in non-line-of-sight, while an IMU provides high-rate motion data but drifts when left alone. Fused, the IMU carries the pose smoothly between UWB fixes and through radio dropouts, and UWB continuously corrects the IMU's drift. The result is a position that is both smooth and anchored - which neither sensor delivers by itself.

Q: How does sensor fusion handle NLOS or losing anchors?

A well-designed fusion engine detects when a UWB measurement is degraded - non-line-of-sight, multipath, or a failed anchor - and reduces its weight or rejects it, letting the IMU and odometry carry the pose until good measurements return. This is where needCode's ranging-layer depth matters: because we build the UWB layer itself, the filter receives measurement-quality context rather than guessing. A filter that trusts a bad range in NLOS is worse than no fusion at all.

Q: What algorithms are used for sensor fusion?

The core is state estimation - most commonly Kalman filter variants (EKF/UKF), alongside complementary filters for simpler cases, particle filters for hard non-linear problems, and ML-based methods where patterns beat models. The right choice depends on the platform's dynamics, compute budget, and accuracy target. needCode selects and implements the method per project rather than forcing one filter onto every problem.

Q: Can sensor fusion run on a microcontroller?

Yes - and designing for that constraint is a needCode specialty: we optimise fusion algorithms for microcontroller-class targets with limited memory and hard power budgets, using efficient math and power-aware scheduling. Fusion that only runs on a workstation is a demo, not a product. This is what makes fused navigation viable on battery-powered tags, robots, and devices.

Q: Does sensor fusion work without GPS?

Yes - indoor and GPS-denied navigation is precisely what UWB-anchored fusion is for: UWB provides the absolute reference satellites can't deliver indoors, and inertial and odometry sensors carry the pose through radio gaps. The same architecture serves warehouses, autonomous robots, and environments where GPS is jammed or spoofed. GNSS is fused in as one more sensor where it's available, not depended on.

Q: How accurate is fused positioning for a moving platform?

Fused UWB positioning delivers centimetre-level accuracy under good conditions, and fusion is what keeps accuracy usable in motion and through disturbances - but the honest answer is that real-world accuracy depends on anchor geometry, environment, platform dynamics, and update rate. That is why needCode measures achievable accuracy on your platform in your environment during a feasibility study, rather than quoting a datasheet figure. Fusion narrows the gap between the datasheet and the site; it doesn't repeal it.

Q: What sensors can be fused together?

Any sensor with characterised behaviour can join the estimate - typical sets combine UWB ranging, MEMS IMUs (accelerometer, gyroscope, magnetometer), LiDAR or visual odometry, wheel odometry, barometers, and GNSS where available; UWB can also contribute radar-mode presence and motion data. The craft is choosing sensors whose failure modes don't overlap and characterising each one honestly. needCode consults on the sensor set as the first step of every fusion engagement.

Q: What's the difference between sensor fusion and an RTLS?

An RTLS (real-time location system) is facility infrastructure that locates many tagged assets and people; sensor fusion is the discipline that keeps one platform's pose solid, especially in motion and through sensor failures. They meet in the middle: fusion makes an RTLS resilient, and an RTLS gives a fused platform its absolute reference. needCode builds both - and they're deliberately separate pages because they answer different buying questions.

Q: Can needCode design a sensor fusion system end to end?

Yes - needCode covers the full chain: sensor selection and trade-off analysis, the UWB ranging layer, fusion algorithm design and implementation on your target hardware, time synchronisation and calibration, HIL and field validation, and integration into the RTLS or operational systems the pose feeds. As a certified Qorvo partner with the largest dedicated UWB team in Central Europe, we work from the radio up. One team owns sensor, algorithm, and integration.

UWB drops in non-line-of-sight. An IMU drifts within minutes. Cameras and LiDAR fail in dust, darkness, and featureless space. GPS ends at the door - and can be jammed. needCode builds navigation-grade sensor fusion that combines them into one pose your system can trust: position, velocity, and orientation that hold through dropouts, drift, and motion - engineered to run on embedded hardware, and tuned in your environment. From an EU-based, certified Qorvo partner.

Book a discovery call

How navigation-grade fusion works

We work with Industry Leaders

One sensor is a single point of failure. Navigation can't afford one.

Every positioning sensor has a blind spot, and the real world finds it. Radio ranging degrades behind metal and people; inertial sensors drift the moment they're left alone; visual odometry dies in the dark and in fog; satellites don't reach indoors and can be denied. A system that navigates on one source works right up until its blind spot - then it doesn't, usually at the worst moment.

Sensor fusion is the engineering answer: combining sensors whose failure modes don't overlap, and - the part that separates working fusion from a textbook filter - knowing how much to trust each measurement at each moment. needCode builds that from the radio up. Because we engineer the UWB ranging layer itself, our filters are fed measurement quality, not just measurements: NLOS conditions, first-path confidence, anchor health. And because we come from embedded, the result runs on microcontrollers within real power budgets - validated hardware-in-the-loop and calibrated in the field, not only in simulation.

Continuous pose

Position, velocity, and orientation that hold through NLOS, dropouts, drift, and sensor faults - in motion, not just at standstill.

From the radio up

We build the UWB ranging layer itself, so the filter knows how much to trust every measurement it's given.

Embedded-grade

Fusion engineered for microcontrollers and real power budgets - HIL-validated, field-calibrated, production-ready.

How navigation-grade sensor fusion works

Working fusion is four disciplines, not one algorithm. We engineer all four - which is why the result survives contact with the real world.

The Sensor Set

Choosing sensors whose failure modes don't overlap: UWB ranging for absolute position, IMU for high-rate motion, LiDAR or visual odometry for structure, wheel odometry and GNSS where they exist. The set is a design decision made against your platform and environment - not a default stack bolted together.

Ultra-Wideband

Sensor Devices

The Estimation Engine

State estimation - Kalman, complementary, and ML-based methods - fuses the streams into one statistically grounded pose, weighting each measurement by how much it deserves to be trusted right now. This weighting is where fusion is won or lost: a filter that trusts a bad UWB range in NLOS is worse than no fusion at all.

Indoor Positioning (RTLS)

Time, Calibration & Truth

Sensors disagree by default - unsynchronised clocks, uncalibrated mounting offsets, and noisy signals turn good hardware into bad data. We engineer time synchronisation and extrinsic calibration, then validate the fused output hardware-in-the-loop and in the field, because accuracy claimed in simulation is a hypothesis, not a result.

Automated Testing & CI/CD

Context & Awareness

Beyond pose, fusion turns where plus what it's doing into operational meaning: a tool used at the correct workstation at the proper angle, an impact on a moving asset, presence detected by UWB radar where cameras can't go. The same engine that navigates a robot can verify a process.

Child Presence Detection

UWB Data Fusion

Building a platform that has to know where it is?

Book a discovery call with our CEO

Book a discovery call

Where sensor fusion goes

The same fusion foundation carries very different platforms. We've built the wireless and embedded depth each one needs.

Autonomous Robots & Humanoids

The pose a robot's autonomy stands on - UWB fused with the robot's own IMU and LiDAR odometry, so localisation holds when light changes, features vanish, and radio drops behind machinery.

Robotics & Humanoids

GPS-Denied Navigation

Navigation that keeps working when satellites are jammed, spoofed, or simply absent - underground, inside structures, in contested environments - from an EU-based partner where data handling and sovereignty matter.

Defence & Aerospace

AGVs, AMRs & Industrial Vehicles

Reliable pose for forklifts, tugs, and mobile robots in metal-dense facilities - the environments where single-source positioning collapses, and where a vehicle that loses itself becomes a safety problem.

Industrial IoT & RTLS

Vehicle Tracking

Contextual Tracking & Process Intelligence

Location fused with motion and environmental data turns tracking into verification - which tool was used where, at what orientation, with what handling - feeding quality, safety, and process systems.

Asset Tracking

Warehouse Positioning

Why fusion projects come to needCode

Fusion from the ranging layer up

As the largest dedicated UWB team in Central Europe and a certified Qorvo partner with nine hardware-platform bring-ups, we don't fuse a black box - we build the ranging layer itself, so measurement quality flows into the filter instead of being guessed.

Embedded-grade, not workstation-grade

We optimise fusion algorithms to run on microcontrollers with limited memory and hard power budgets - the constraint most fusion work ignores until the battery target fails. Fusion that only runs on a dev laptop isn't a product.

Validated where it counts

Calibration procedures, hardware-in-the-loop rigs, and field tuning against your platform in your environment - plus our own UWB protocol sniffer to see what the radio is really doing when results drift.

The whole system, one team

From custom sensor hardware and edge processing to the RTLS and operational systems the pose feeds - one team owns sensor, algorithm, and integration, so accuracy doesn't get lost in a handover.

Four ways to bring needCode in

From a sensor-architecture study to a standing team. We match the engagement to where the platform is.

01 Sensor Architecture & Feasibility Study

Duration:
4-6 weeks
Best for:
Choosing the sensor set and fusion approach for your platform - and measuring what accuracy is actually achievable in your environment before committing
Deliverable:
Sensor selection and trade-off analysis, fusion architecture, measured feasibility results, leadership readout

02 Fusion Stack Build

Duration:
Phased
Best for:
Implementing the estimation engine on your target hardware - filters, calibration, time sync, and the UWB ranging layer beneath them
Deliverable:
A working fusion stack on target silicon, calibration procedures, test harness, documentation

03 Field Tuning & Hardening

Duration:
Phased
Best for:
Taking fusion that works in the lab to fusion that survives the site - NLOS handling, outlier rejection, drift management, validation at operational tempo
Deliverable:
Field-calibrated system, measured performance report, hardened failure handling

04 Embedded Team

Duration:
Multi-year, retainer-based
Best for:
Platform companies who want a dedicated positioning-and-fusion squad inside their programme
Deliverable:
An embedded team in your cadence - the model behind the 30-FTE Qorvo programme

What we ship on

We pick the sensor set and estimation method that match your platform, environment, and power budget.

Sensors

UWB (ranging + radar)

wheel odometry

LiDAR & visual odometry

barometer

GNSS (where available)

MEMS IMU (accelerometer / gyroscope / magnetometer)

Estimation

Kalman filters (EKF / UKF)

complementary filters

particle filters

NLOS & outlier rejection

ML-based fusion

Embedded

Microcontroller-class targets

Zephyr

FreeRTOS

optimised math for constrained platforms

power-aware scheduling

Validation

HIL test rigs

field calibration procedures

UWB Protocol Sniffer (in-house)

Silicon

Qorvo QM33 / QM35

Nordic (BLE companion)

legacy Decawave DW1000 / DW3000

Integration

RTLS / location engines

MQTT / REST

digital twin feeds

ROS 2

Case studies

needCode doesn't publish a named navigation-fusion deployment. What we can show is the ranging, sensing, and embedded depth that fusion stands on - and where it's already at work.

Qorvo: RF Leadership

Context: Rapid scaling for new chipset bring-up.

Scale: Grew from <10 to 30 FTEs.
Output: Supported bring-up of 9 new hardware platforms (SDKs, Drivers, Stacks).
Retention: Zero-churn core team retained for 5+ years.

Dedicated Development Center for RF Solutions

Bluetooth Mesh Smart Lighting Control System

Smart Lighting: Core R&D Extension

Context: Client needed deep, specialized expertise to pivot from proprietary tech to a new global standard.

Service: Deployed a dedicated squad of embedded engineers to function as the client's core R&D team.
Output: Co-authored official Bluetooth SIG protocols and delivered the world’s first certified BLE Mesh stack.
Value: Enabled the client to secure Series A funding and defined the industry standard for smart buildings.

Creative Werks: Innovation rescue

Context: Hardware obsolescence threatened production shutdown.

Action: Full-stack takeover (PCB redesign + Firmware + Mobile App).
ROI: 1230% ($1.6M value generated).
Speed: Payback period of 2-3 months.

NeedCode-case study - IoT Solution for Boat Lift Modernization - cover2s

PowerPollen: AgTech automation

Context: Lack of internal expertise stalled a critical automation project.

Action: Re-architected system using unified MCU and ISOBUS standards.
ROI: 13.8x ($2.9M value generated).
Impact: Enabled $1.9M increase in harvester value.

Strategic Partnership

needCode is an official business partner of Qorvo, bringing over 8 years of proven expertise and trusted service to the technology sector.

Check our performance

Members of the UWB Alliance

In 2025 we became a member of the UWB Alliance. This strategic step reinforces our commitment to pioneering Ultra-Wideband (UWB) technology.

Proudly Certified for Excellence and Security

needCode is officially certified for:
ISO 9001:2015 – Quality Management
ISO/IEC 27001:2022 – Information Security

ISO certifications reflect our focus on delivering reliable IoT solutions, smart product development, and secure technology services.

Testimonials

“I think the key takeaway from needCode is their ability to adapt and understand the customer's requirements. That took away probably a large portion of what could have been a lot of development time and expense for both companies.”

Bob Folkestad

President at Creative Werks

“One aspect that truly sets needCode apart is its profound expertise in firmware development. Their proficiency in various programming languages, embedded systems and hardware architecture is truly impressive. When faced with difficult problems, their strong problem-solving skills and analytical mindset shine through, allowing them to overcome obstacles with remarkable ease.”

Semeh Sarhan

CEO at Xtrava

“I worked with needCode while leading the NWTN-Berlin team in 2018. A big chunk for our FW development has been outsourced to them and they had proven to iterate very quickly, following specs and deliver on time. It was great working with them. I recommend working with needCode’s team on any Embedded SW development.”

Marco Salvioli Mariani

CTO at NWTN Berlin GmbH

“needCode Team proved to be one of the best engineers I have ever met. The part I like the most about the team is the more difficult an obstacle seems to be, the more motivated they were to find a solution and a way forward.”

Szymon Słupik

CTO at Silvair

“needCode is an outstanding partner. Their quick follow-up, scalability, and extensive professional network set them apart. Their expertise in wireless technologies has been valuable, supporting us from low-level drivers to architecture discussions.”

Tim Allemeersch

Director at Qorvo, Inc.

“needCode did a great job improving the firmware of the Vai Kai connected toys and developing new features, surpassing our expectations multiple times. I would definitely recommend hiring Bartek and needCode for the embedded software projects!”

Matas Petrikas

CEO & Co-founder at Vai Kai UG

Insights

UWB Data Fusion: Combining IMUs & Sensors for Contextual RTLS

Knowing an asset’s location is only half the story. A dot moving on a map provides no operational context. It cannot tell you if a worker has fallen, if a vehicle is being used productively, or if a critical process is being performed correctly. To unlock the next level of operational intelligence, we must evolve […]

Basics of UWB indoor location systems (indoor RTLS)

Introduction to real-time location systems (RTLS) using ultra-wideband technology RTLS stands for Real Time Location System. As you may guess, such a system aims to provide real-time information about the location of “an object” with respect to “a coordinate system”. This loose definition raises three questions: – Provide information to whom? – What is “an […]

FAQ