60 UWB Indoor Positioning Systems
Sammy the Sensor wanted to build an indoor positioning system: “GPS does not work inside buildings – how do I find my way around a warehouse?” Max the Microcontroller explained: “We put UWB anchors on the walls and ceiling, like lighthouses in a harbor. Each anchor knows its own position. When a tag sends a signal, the anchors measure how long it took to arrive, and from those time measurements, we can calculate exactly where the tag is!” Bella the Battery asked: “How many lighthouses do we need?” Max replied: “At least 3 for flat positioning, 4 for 3D. But the important rule is: NEVER put them all in a line! Spread them out in a square or triangle for best accuracy.” Lila the LED warned: “And watch out for metal shelves – they block signals! Mount anchors high up so they can see over obstacles.”
60.2 Learning Objectives
By the end of this chapter, you will be able to:
- Design UWB-based indoor positioning systems with appropriate anchor placement and geometric diversity
- Calculate anchor density and placement geometry for specific accuracy requirements using GDOP analysis
- Differentiate TWR-based and TDoA-based system architectures based on scale and infrastructure constraints
- Evaluate commercial UWB chipsets and development platforms for specific deployment scenarios
- Analyze how GDOP (Geometric Dilution of Precision) amplifies ranging errors into position errors
UWB (Ultra-Wideband) can locate objects indoors with centimeter-level accuracy, which GPS cannot do. By measuring the precise time radio pulses take to travel between anchors and a tag, UWB calculates exact positions. This enables applications like tracking tools in a warehouse or guiding robots through a factory.
60.3 Introduction
Designing a UWB positioning system requires careful consideration of anchor placement, ranging method, and system architecture. This chapter covers the engineering principles and practical considerations for deploying UWB infrastructure.
Anchor placement is critical: Never place anchors in a line. Good geometric diversity (spread across the space, varying heights) dramatically improves position accuracy through better GDOP.
60.4 Anchor Placement Principles
Minimum Requirements:
- 2D positioning: 3 anchors (trilateration in plane)
- 3D positioning: 4 anchors (trilateration in space)
- Practical systems: 6-8 anchors for redundancy and coverage
Geometric Diversity Rules:
- Avoid Collinearity: Never place all anchors in a straight line
- Height Variation: For 3D, vary anchor heights (walls, ceiling)
- Coverage Overlap: Each point should see 4+ anchors
- DOP Consideration: Good geometric dilution of precision
Typical Deployment Densities:
| Environment | Area per Anchor | Anchor Height | Configuration |
|---|---|---|---|
| Office | 200-300 m² | 2.5-3 m (ceiling) | Grid pattern |
| Warehouse | 300-400 m² | 4-8 m (high ceiling) | Grid pattern |
| Manufacturing | 150-250 m² | 3-5 m (overhead) | Process flow |
| Retail | 100-200 m² | 2.5-3 m | Customer paths |
| Hospital | 150-250 m² | 2.5-3 m | Room + corridor |
60.5 Line-of-Sight Considerations
UWB performs best with clear line-of-sight but can tolerate some obstructions:
- Direct LOS: Best performance (10-20 cm accuracy)
- NLOS Soft (drywall, glass): Degraded (30-50 cm accuracy)
- NLOS Hard (metal, water): Significant error or no signal
- Multipath: Can cause errors, mitigated by antenna design
Mitigation Strategies:
- Place anchors high to maximize LOS
- Use multiple anchors for redundancy
- NLOS detection algorithms
- Kalman filtering to smooth positions
60.6 Worked Example: UWB Anchor Placement for Warehouse Asset Tracking
Scenario: A logistics company needs to track forklifts with 30cm accuracy in a 60m x 40m warehouse with 8-meter high ceilings and metal shelving racks.
Given:
- Warehouse area: 2,400 m^2 (60m x 40m)
- Ceiling height: 8 meters
- Metal shelving racks: 6 meters tall, arranged in 4 parallel rows
- Required accuracy: 30 cm (sub-meter for collision avoidance)
- Tracking targets: 12 forklifts with active UWB tags
- Budget constraint: Minimize anchor count while meeting accuracy
Steps:
- Calculate minimum anchor density:
- For 30cm accuracy with TDoA, need GDOP < 2.0
- Good GDOP requires anchor spacing of 15-20 meters
- Area per anchor: 225-400 m²
- Minimum anchors: 2,400 / 300 = 8 anchors (baseline)
- Account for shelving obstructions:
- Metal shelves block UWB signals (20+ dB attenuation)
- Each tracking point needs LOS to 4+ anchors
- Add 50% redundancy for NLOS mitigation: 8 x 1.5 = 12 anchors
- Determine anchor positions:
- Perimeter anchors at 15m spacing: 8 anchors at (0,0), (0,20), (0,40), (30,0), (30,40), (60,0), (60,20), (60,40)
- Interior anchors between shelving rows: 4 anchors at (15,13), (15,27), (45,13), (45,27)
- Total: 12 anchors
- Optimize anchor height:
- Mount at 7 meters (below 8m ceiling)
- This provides LOS over 6m shelving with 30-degree clearance angle
- Calculation: tan(30) x 7m = 4m horizontal clearance from shelf top
- Verify coverage with geometry analysis:
- Simulate GDOP across warehouse floor
- Worst case GDOP: 1.8 (in corners between shelves)
- Position accuracy: 15cm UERE x 1.8 GDOP = 27cm (meets 30cm requirement)
GDOP (Geometric Dilution of Precision) amplifies ranging errors into position errors. The relationship is:
\[\sigma_{position} = GDOP \times \sigma_{ranging}\]
Example: With UWB ranging uncertainty \(\sigma_{ranging} = 15\) cm and GDOP = 1.8, position uncertainty is \(\sigma_{position} = 1.8 \times 15 = 27\) cm. If anchors were collinear (GDOP ≈ 10), the same 15 cm ranging error would yield \(10 \times 15 = 150\) cm position error—6× worse. This is why anchor geometry matters more than ranging precision. Good anchor placement (GDOP < 2) is worth 10× more than improving ranging precision from 15 cm to 10 cm.
Result: Deploy 12 UWB anchors at 7-meter height achieving 27cm average positioning accuracy across the 2,400 m² warehouse floor.
Key Insight: Anchor height is critical in warehouses with tall shelving. Mounting anchors at 70-90% of ceiling height (7m in 8m space) provides clearance over obstacles while maintaining good GDOP geometry. If anchors were mounted at ceiling level (8m), the near-vertical geometry would degrade horizontal accuracy by 40%.
60.7 System Architecture Options
UWB positioning systems can be architected in several ways depending on requirements.
60.7.1 TWR-Based Architecture (Simple, Small Scale)
- Tags: Active devices performing ranging with anchors
- Anchors: Fixed positions, respond to tag polls
- Gateway: Collects ranging data via wired/wireless backhaul
- Location Server: Computes positions from ranges
- Application Layer: Visualization, alerts, analytics
Pros: Simple, no anchor synchronization, lower infrastructure cost Cons: Tags consume more power, scales to ~50-100 tags
60.7.2 TDoA-Based Architecture (Complex, Large Scale)
- Tags: Passive beacons, only transmit blinks
- Anchor Network: Synchronized via wired Ethernet or GPS
- Positioning Engine: Centralized or distributed, performs trilateration
- RTLS Platform: Real-time location services, geofencing, analytics
- Integration Layer: APIs for third-party systems
Pros: Scales to 1000s of tags, low tag power, high update rate Cons: Complex infrastructure, anchor sync critical, higher cost
60.7.3 Hybrid and Peer-to-Peer Modes
Modern systems often combine approaches:
- TWR for configuration: Anchors use TWR to self-survey positions
- TDoA for tracking: Tags use TDoA for power efficiency
- Peer-to-peer ranging: Device-to-device ranging without infrastructure (e.g., Apple U1 chip)
60.7.4 Why TDoA Requires Wired Synchronization (and What Happens Without It)
The accuracy of TDoA depends on anchors measuring the difference in arrival times of a tag’s blink with sub-nanosecond precision. One nanosecond of timing error translates to 30 cm of position error (since radio signals travel at the speed of light: 3 x 108 m/s x 1 ns = 0.3 m). This means anchors must share a common time reference accurate to within approximately 0.3 ns to achieve 10 cm positioning.
In practice, this synchronization is achieved through one of three methods:
Wired Ethernet with PTP (IEEE 1588): Most common in enterprise deployments. PTP over Ethernet achieves sub-microsecond synchronization, and UWB-specific extensions (using wireless calibration packets between anchors) reduce residual error to under 1 ns. This is why most TDoA systems (Ubisense, Sewio, Kinexon) require Ethernet cabling to every anchor.
Wireless anchor-to-anchor ranging: Some systems (Decawave MDEK1001) use TWR between anchor pairs to estimate clock offsets. This eliminates wiring but introduces 1-3 ns of synchronization jitter, degrading position accuracy from 10 cm to 30-50 cm. Acceptable for asset tracking but insufficient for safety-critical applications.
GPS-disciplined clocks: Outdoor-capable anchors can use GPS timing (10-20 ns accuracy). This works for logistics yards but is impractical indoors where GPS signals are unavailable.
The cost implication is significant: A TDoA system for a 2,400 m2 warehouse requires Ethernet to 12 anchor points. At approximately $150/run for structured cabling, this adds $1,800 in infrastructure cost beyond the anchors themselves. A TWR system eliminates this cabling cost entirely but cannot scale beyond approximately 50 simultaneous tags due to airtime contention (each tag must exchange multiple packets with each anchor).
60.8 UWB vs Other Technologies
Understanding when to use UWB requires comparing it to alternative positioning technologies.
60.8.1 Comprehensive Technology Comparison
| Technology | Accuracy | Range | Infrastructure | Relative Cost | Best Use |
|---|---|---|---|---|---|
| GPS | 3-5 m | Global | Satellites | Low | Outdoor navigation |
| Wi-Fi RSSI | 3-5 m | 50 m | Existing APs | Low | Room-level indoor |
| Wi-Fi RTT | 1-2 m | 50 m | Wi-Fi 6 APs | Medium | Indoor navigation |
| BLE AoA | 0.5-1 m | 30 m | BLE 5.1 anchors | Medium | Asset tracking |
| UWB | 0.1-0.3 m | 70 m | UWB anchors | Medium | Precision positioning |
| Vision | 0.01-0.1 m | 10 m | Cameras + compute | High | Robotics and AR/VR |
60.8.2 When to Choose UWB
UWB is ideal when:
- Sub-meter accuracy is required (automotive, industrial safety)
- High update rates needed (real-time tracking)
- Secure ranging required (access control)
- Line-of-sight is generally available
- Infrastructure cost is acceptable
Consider alternatives when:
- Room-level accuracy sufficient (Wi-Fi/BLE cheaper)
- Outdoor use required (GPS better)
- Extreme range needed (>100m)
- Ultra-low cost critical (BLE beacons)
- No line-of-sight available (vision-based may be better)
60.9 Hardware Platforms
The UWB ecosystem has matured significantly with several commercial chipsets and development platforms.
60.9.1 Commercial Chipsets
Qorvo/Decawave DW3000 Series:
- Most widely deployed UWB chipset
- IEEE 802.15.4z compliant
- Channels 5 and 9 (6.5 GHz, 8 GHz)
- ~10 cm ranging accuracy
- Integrated MAC and PHY
- Low power: 60 mW in active mode
- Used in: Industrial tracking, automotive, consumer
NXP Trimension Family:
- SR040 (automotive-grade)
- SR150 (mobile/consumer)
- SR100T (secure element integrated)
- IEEE 802.15.4z HRP UWB
- Enhanced security features
- Used in: BMW digital keys, Samsung phones, Apple iPhones (rumored)
Apple U1 / U2 Chip:
- Proprietary design (likely based on licensed IP)
- Integrated in iPhone 11+ (U1), iPhone 15+ (U2)
- Also in AirTags, HomePod mini, Apple Watch Ultra
- ~50 billion potential devices
- Closed ecosystem but industry-defining
STMicroelectronics SR21:
- Automotive-qualified
- Integrated Arm Cortex-M33
- Secure ranging with cryptography
- Target: Digital car keys
60.9.2 Development Kits and Platforms
Qorvo DWM3000EVB:
- Evaluation board for DW3000
- Arduino-compatible headers
- Reference TWR and TDoA firmware
- ~$200 per board
- Best for: Prototyping, academic research
Decawave MDEK1001:
- Complete development kit (12 modules)
- Pre-configured for TDoA and TWR
- Android app for visualization
- Out-of-box positioning demo
- ~$1500 for kit
- Best for: Quick proof-of-concept
Pozyx Creator:
- All-in-one positioning system
- Hardware + software + API
- Anchors, tags, gateway included
- Cloud dashboard
- ~$2000+ for starter kit
- Best for: Enterprise deployment, minimal development
60.9.3 Custom PCB Integration
Key considerations when integrating UWB chips:
- Antenna design: Critical for performance (typically PCB trace or chip antenna)
- RF layout: Follow reference designs carefully
- Power supply: Clean, low-noise power required
- Crystal: High-quality oscillator for timing accuracy
- Regulatory: FCC/CE certification required
60.10 Knowledge Check
Common Pitfalls
Anchors on one wall produce poor geometry (high GDOP) for tags on the opposite side. Fix: distribute anchors to maximise spatial diversity — anchors should surround the coverage area, not line one wall.
Metal shelving, machinery, and moving forklifts create NLOS conditions that degrade UWB accuracy to 50–100 cm. Fix: set realistic accuracy expectations for the specific environment and use additional anchors to reduce NLOS probability.
TDOA positioning requires nanosecond-level synchronisation between all anchors, which demands a dedicated wired timing network or ultra-stable clock distribution. Fix: evaluate the synchronisation infrastructure cost alongside anchor hardware cost when planning a TDOA-based UWB RTLS.
60.11 Summary
UWB positioning system design requires balancing accuracy requirements with infrastructure complexity and cost. Proper anchor placement with good geometric diversity is essential for achieving centimeter-level accuracy.
Key Takeaways:
Anchor Placement: Avoid collinearity, vary heights for 3D, ensure 4+ anchor visibility per tracked point
System Architecture: TWR for simple/small deployments, TDoA for enterprise scale
Environment Matters: Line-of-sight preferred, metal and water cause significant degradation
Technology Selection: UWB excels at sub-meter accuracy, high update rates, and secure ranging
Commercial Options: Mature ecosystem with Qorvo DW3000, NXP Trimension, Apple U1/U2 chipsets
60.12 Concept Relationships
Understanding how UWB positioning systems relate to other concepts:
Builds Upon:
- UWB Ranging Techniques: Positioning systems apply TWR and TDoA ranging methods to calculate device locations
- TDMA Scheduling: Similar time-division concepts for collision-free communication (WirelessHART comparison)
Enables:
- UWB Applications and Security: Real-world deployments of UWB positioning in automotive, industrial, and consumer contexts
- Asset Tracking: UWB provides centimeter-level tracking for industrial and logistics applications
Alternatives and Comparisons:
- Wi-Fi RTT (802.11mc): 1-2m accuracy, lower cost, but less precise than UWB
- BLE AoA (Angle of Arrival): 0.5-1m accuracy, Bluetooth ecosystem compatibility
- Vision-based systems: Sub-centimeter accuracy but high computational cost
Key Insight: UWB positioning excels when sub-meter accuracy is essential and infrastructure cost is acceptable. For room-level accuracy, Wi-Fi or BLE are more cost-effective alternatives.
60.13 See Also
Related Protocols:
- Bluetooth BLE: Proximity-based positioning using RSSI fingerprinting
- Wi-Fi Positioning: IEEE 802.11mc RTT for meter-level indoor positioning
- Zigbee Mesh: Mesh topology for sensor networks (non-positioning)
Commercial Platforms:
- Pozyx Creator: All-in-one UWB positioning platform with cloud dashboard and API
- Decawave (Qorvo) MDEK1001: Development kit for TDoA and TWR experimentation
- Apple U1/U2 Chip: Consumer UWB in iPhones, AirTags, and HomePod mini
Standards and Specifications:
- IEEE 802.15.4z: UWB High Rate Pulse (HRP) standard with enhanced security
- FiRa Consortium: Industry alliance for UWB interoperability and secure ranging
Further Reading:
- Qorvo DW3000 User Manual: TWR and TDoA implementation details
- IEEE 802.15.4z-2020 Standard: Enhanced Impulse Radio UWB specification
60.14 Try It Yourself
60.14.1 Challenge 1: Design a Warehouse Positioning System
Scenario: A logistics warehouse (100m × 60m, 8m ceiling height) needs to track 200 forklifts with 30cm accuracy.
Your Tasks:
- Calculate minimum anchor count for 30cm accuracy using TDoA
- Plan anchor placement avoiding metal shelving NLOS
- Determine optimal anchor height for LOS over 6m obstacles
- Validate GDOP < 2.0 across the floor
Given:
- UERE (User Equivalent Range Error) = 15cm for UWB TDoA
- Target accuracy: 30cm (95th percentile)
- Metal shelving: 6m tall, 20+ dB attenuation
Hint: Use GDOP formula: Position Accuracy = UERE × GDOP. Solve for GDOP needed, then plan anchor geometry.
Solution
- GDOP calculation: Target 30cm = 15cm × GDOP → GDOP < 2.0 required
- Anchor density: 300-400 m² per anchor (industrial environment) → 6,000 m² / 300 = 20 anchors baseline
- NLOS mitigation: Add 50% for metal shelving → 20 × 1.5 = 30 anchors
- Anchor placement: Perimeter + interior grid at 7m height (above 6m shelves)
- Validation: Simulate GDOP at 10m × 10m grid points, verify all < 2.0
60.14.2 Challenge 2: TWR vs TDoA Trade-off Analysis
Scenario: Smart home with 15 UWB tags (people, pets, robots) across 200 m².
Compare TWR and TDoA:
- Calculate tag battery life for 1 Hz position updates (TWR vs TDoA)
- Determine infrastructure cost (anchors, gateways, sync hardware)
- Evaluate scalability to 50 tags
Given:
- TWR: Tag transmits 4× per update (poll 4 anchors), 10 mW × 3 ms per transmission
- TDoA: Tag transmits 1× per update (blink), 10 mW × 1 ms, but needs 4+ synchronized anchors
Hint: TWR uses more tag power but simpler infrastructure. TDoA is opposite.
Solution
TWR Analysis:
- Tag energy per update: 4 transmissions × 10 mW × 3 ms = 120 µJ
- Average power at 1 Hz: 120 µJ/s = 0.12 mW
- Battery (CR2032, 220 mAh @ 3V = 2,376 J): 2,376 J / 0.00012 W = 19.8M seconds = 229 days
- Infrastructure: 4 anchors (~$200 each), no sync needed = $800
TDoA Analysis:
- Tag energy per update: 1 transmission × 10 mW × 1 ms = 10 µJ
- Average power at 1 Hz: 10 µJ/s = 0.01 mW
- Battery life: 2,376 J / 0.00001 W = 237.6M seconds = 7.5 years (75× longer!)
- Infrastructure: 4 anchors + Ethernet + PTP synchronization = $800 + $500 cabling = $1,300
Verdict: For 15 tags in a smart home, TWR is simpler (no wiring needed) with adequate battery life (~229 days). TDoA offers dramatically longer battery life (~7.5 years) but requires synchronized anchor infrastructure ($500+ cabling). For 50+ tags, TDoA’s scalability and power advantages are decisive.
60.14.3 Challenge 3: Anchor Failure Impact
Scenario: UWB positioning system with 8 anchors in a 60m × 40m warehouse. One anchor fails.
Analyze:
- Which devices lose positioning capability completely?
- How does GDOP change in affected zones?
- What is the failover strategy?
Hint: Check graph routing redundancy and GDOP degradation with N-1 anchors.
Solution
- Coverage loss: Devices relying on failed anchor as primary route lose positioning. Devices with 3+ remaining anchors retain positioning (TDoA minimum).
- GDOP degradation: Zones with 4→3 anchors see GDOP increase from ~1.5 to ~2.5 (67% worse). Accuracy degrades from 15cm to 25cm.
- Failover: Network Manager detects missing anchor (no heartbeat), recalculates routes using remaining 7 anchors, tags automatically use backup graphs. Recovery time: <1 second.
Prevention: Deploy 12 anchors instead of 8 (50% redundancy). Every zone has 4+ anchors even after 1-2 failures.
60.15 What’s Next?
| Chapter | Description |
|---|---|
| UWB Applications and Security | Real-world UWB deployments in automotive, industrial, and consumer contexts with security features |
| UWB Ranging Techniques | TWR, TDoA, and AoA ranging methods that underpin positioning calculations |
| UWB Fundamentals | Core UWB radio technology including impulse radio, bandwidth, and IEEE 802.15.4z |
| NFC Fundamentals | Near-field communication as a complementary short-range identification technology |
| RFID Design and Deployment | Deployment planning parallels for RFID systems in similar environments |