UWB-Based Autonomous Following Technology: Principles and Core Architecture

1. Introduction

With the rapid growth of service robots, autonomous vehicles, smart luggage, and mobility-assist devices, autonomous following has emerged as a key interaction paradigm. Traditional methods rely heavily on vision-based approaches (e.g., object detection, color tracking) or wireless signal strength estimation (e.g., Bluetooth RSSI, Wi-Fi RTT). However, these approaches often face limitations in accuracy, latency, and stability.

Ultra-Wideband (UWB) technology introduces a breakthrough in this field. With centimeter-level accuracy, sub-millisecond latency, and strong resistance to multipath interference, UWB enables robust tracking and following in both indoor and outdoor environments.

2. Technical Principles

The foundation of UWB following lies in high-precision, real-time localization. UWB transmits extremely short pulses across ultra-wide frequency bands (hundreds of MHz to several GHz) and calculates relative position using Time of Flight (TOF) or Time Difference of Arrival (TDOA).

2.1 Signal Characteristics

Pulse Width: < 1 ns
Bandwidth: > 500 MHz (FCC standard: 3.1–10.6 GHz)
Penetration: Can pass through non-metallic materials such as wood, plastic, and the human body
Multipath Resistance: Narrow pulses minimize overlap of reflected signals, improving accuracy

2.2 Localization Methods

TOF (Time of Flight):
Measures round-trip travel time of a signal to compute distance:

d = c \times \left( t_{round} - t_{proc} \right) / 2

where c is the speed of light, t_roundis round-trip time, and t_proc is processing delay.

TDOA (Time Difference of Arrival):
Multiple anchors receive a tag’s signal, and differences in arrival times are used to triangulate position.
PDOA (Phase Difference of Arrival):
Measures phase differences of the same signal across multiple antennas. This enables Angle of Arrival (AoA) estimation and improves direction-aware positioning.

2.3 System Architecture

A typical UWB autonomous following system includes:

Tag: Worn by the user (e.g., on a belt or backpack), periodically transmitting signals
Anchors: Mounted on the follower device (robot, wheelchair, luggage) or fixed in the environment
Main Controller (MCU): Executes localization and path planning algorithms
Drive System: Controls motors based on real-time target position and velocity

3. Engineering Implementation

3.1 Hardware Components

UWB Chipsets: Decawave/Qorvo DW1000 & DW3000 series
MCUs: STM32, ESP32, NXP i.MX RT series
Auxiliary Sensors: IMU (accelerometer, gyroscope), wheel encoders, ultrasonic sensors (collision avoidance)

3.2 Sensor Fusion

To improve robustness, UWB is often combined with other sensors:

UWB + IMU: Uses inertial data to bridge short signal dropouts
UWB + Vision: Visual tracking for close-range precision, UWB for long-range stability
UWB + Ultrasonic: Enhances obstacle avoidance at short distances

Common fusion algorithms: EKF (Extended Kalman Filter), UKF (Unscented Kalman Filter), Particle Filtering for multi-target tracking.

3.3 Path Planning Modes

Direct Following: Straight-line tracking in open spaces
Obstacle-Aware Following: Uses LiDAR or Visual SLAM to reroute dynamically
Queue Following: Multiple units follow in formation

4. Advantages and Challenges

Advantages:

High accuracy (±10 cm)
Low latency (<10 ms)
Strong anti-interference and penetration capability

Challenges:

High deployment cost for large outdoor areas
Accuracy degradation in heavy obstruction (especially metallic structures)
Additional power optimization required for small mobile devices

5. Application Examples

PSICV Logistics Robot: Combines multi-sensor fusion and path planning to follow warehouse operators
Soffofel Smart Wheelchair: Multi-directional following with one-touch recall, enabling caregivers to be tracked seamlessly in malls and hospitals
Soffofel Smart Luggage: Provides omnidirectional autonomous following and one-touch recall in airports

6. Outlook and Future Trends

Looking ahead, UWB-based autonomous following will evolve towards:

Low-power, miniaturized modules suitable for wearables and compact devices
Multi-sensor fusion with vision, LiDAR, and IMU for reliability in all conditions
Cloud-enabled coordination, supporting swarm robotics and shared data environments
Cross-technology integration with Bluetooth LE and Wi-Fi RTT for seamless operation across diverse scenarios

Ultimately, UWB will transform autonomous following into a ubiquitous, natural, and intuitive interaction method across consumer, industrial, and mobility applications.