UWB3000 TDoA, SS-TWR and DS-TWR: A Practical Architecture for High-Accuracy Indoor Localization

Ultra-Wideband (UWB) technology has become one of the most attractive solutions for high-accuracy localization in environments where conventional Global Navigation Satellite System (GNSS) signals are unavailable or unreliable. Indoor robotics, unmanned aerial vehicles, industrial automation, asset tracking, human-rescue systems, and GPS-denied navigation all require localization methods capable of providing short-range distance measurements with low latency and centimeter-level potential accuracy. This work explores a practical experimental platform based on UWB3000/DWM3000-class transceivers and ESP32-S3 controllers, combining three important ranging and localization techniques:

• Single-Sided Two-Way Ranging (SS-TWR)

• Double-Sided Two-Way Ranging (DS-TWR)

• Time Difference of Arrival (TDoA)

  
  

The experimental hardware shown above consists of two UWB anchor nodes and one target node. Each UWB module is interfaced with an ESP32-S3 controller for configuration, packet exchange, timestamp acquisition, and communication with a host computer. The platform provides a useful testbed for studying ranging accuracy, synchronization, clock drift, Time-of-Flight estimation, TDoA positioning, and hybrid localization algorithms.

Why UWB for GPS-Denied Localization?

Localization becomes considerably more difficult when GNSS signals are blocked by buildings, tunnels, industrial structures, rubble, vegetation, or underground environments. A drone or mobile robot operating in such an environment requires an additional source of geometric information. UWB is particularly attractive because it transmits extremely short radio pulses over a wide bandwidth. The large bandwidth provides excellent time resolution, enabling the receiver to estimate the radio signal's propagation time. The fundamental principle is simple: a radio signal travels between two devices in a propagation time TTOF , the corresponding distance is d=cTTOF.

where

• dd is the transmitter-to-receiver distance,

• cc is the speed of light

• TTOF is the measured Time of Flight.

• 
  

The difficulty is that radio signals travel extremely quickly. A timing error of only 1 ns corresponds to approximately 0.30 m. Therefore, accurate ranging requires highly precise hardware timestamps and careful compensation for antenna delay, clock drift, multipath, non-line-of-sight propagation, and processing latency.

Experimental UWB Localization Platform

The prototype consists of three principal nodes: Anchor 1 → Target ← Anchor 2

The two anchors are installed at known positions, while the target represents the mobile node whose position must be estimated. Each subsystem contains:

System Architecture

• UWB transceiver→ precise UWB packet transmission and timestamp generation

• ESP32-S3→ UWB configuration, packet control, timestamp processing, USB/serial communication

• Host computer→ distance calculation, TDoA processing, visualization, data logging, filtering, and potentially machine-learning-based correction.

Single-Sided Two-Way Ranging

Single-Sided Two-Way Ranging is one of the simplest UWB ranging techniques. Instead of requiring the anchor and target clocks to be synchronized, one device initiates a packet exchange. The sequence can be represented as

For example,

Why DS-TWR Can Be More Accurate

Consider two UWB nodes containing oscillators that differ slightly in frequency.

causes one device to measure a time interval slightly differently from the other device.

SS-TWR depends significantly on the timing interval measured using one side of the connection.

DS-TWR instead uses timing information from both devices. This allows much of the oscillator error to cancel mathematically. Therefore, DS-TWR generally provides better robustness when:

• anchor and target oscillators have different clock offsets,

• temperature varies,

• packets require relatively long processing delays,

• centimeter-level ranging is required.

  
  

The cost is an extra packet and therefore additional airtime and latency.

SS-TWR Versus DS-TWR

A practical comparison is:

Characteristic	SS-TWR	DS-TWR
Packet exchanges	Fewer	More
Implementation	Simpler	More complex
Measurement latency	Lower	Higher
Channel usage	Lower	Higher
Power consumption	Lower	Slightly higher
Clock-drift sensitivity	Higher	Lower
Expected accuracy	Good	Generally better
Mobile-node application	Very suitable	Very suitable
High-precision experiments	Possible	Preferred

For a swarm containing many nodes, SS-TWR can reduce network traffic. For laboratory characterization or precision positioning, DS-TWR is often preferred.

From Ranging to TDoA

TWR answers a simple question:

How far is the target from an anchor?

TDoA answers a different question:

What is the difference between the target's propagation distance to two receivers?

Important Limitation of a Two-Anchor TDoA System

The experimental system shown contains two anchors. With only a single TDoA measurement from two anchors, the target position generally cannot be uniquely determined in unrestricted two-dimensional space. Instead, TDoA identifies a hyperbolic locus containing multiple possible target locations. For example:

A third synchronized anchor would normally provide another TDoA equation.

Alternatively, the two-anchor system can combine TDoA with:

• DS-TWR ranges,

• SS-TWR ranges,

• known altitude,

• constrained motion,

• IMU information,

• optical flow,

• visual odometry,

• previous target position,

• Kalman filtering,

• drone motion models.

This makes the combination of TDoA + TWR particularly interesting.

Main Experimental Challenges

Several practical factors ultimately determine UWB localization accuracy.

• Antenna orientation: Changing the relative orientation of UWB antennas can modify antenna group delay and received signal strength.

• Multipath: Reflections can distort the detected first path.

• NLOS propagation: Obstacles generally introduce positive range bias.

• Anchor geometry: Poor anchor geometry can amplify position error even when range measurements are accurate.

• Clock drift: Especially important for TDoA.

• Antenna delay calibration: Incorrect TX/RX antenna delay creates a systematic offset.

• USB and PC timing: PC timestamps should not replace the hardware UWB timestamps for Time-of-Flight estimation. USB and operating-system delays are much too large and nondeterministic.

• Update rate: Higher ranging frequency provides faster tracking but increases network utilization.

Conclusion

The UWB3000 two-anchor-one-target platform provides a compact experimental environment for understanding the fundamental mechanisms underlying high-accuracy UWB localization.

SS-TWR offers a simple and communication-efficient method for measuring distance. Its low packet overhead makes it attractive for systems containing many mobile nodes, although oscillator mismatch can introduce measurement errors.

DS-TWR exchanges additional timing information and significantly reduces sensitivity to clock-frequency differences between the anchor and target. For this reason, it is particularly attractive for high-precision ranging and laboratory characterization.

TDoA, in contrast, measures the difference in signal arrival times at multiple anchors. It can support highly scalable localization because the mobile target may only need to transmit a single UWB message. Its principal technical challenge is precise synchronization between the anchors.

For the two-anchor prototype, TDoA alone defines a hyperbolic target-location constraint rather than a unique unrestricted 2D position. Combining TDoA with SS/DS-TWR ranging, previous motion information, IMU measurements, visual odometry, or additional anchors provides a much stronger localization solution.

---