# On the Characterisation of the Time-of-Flight VL53L5CX Sensor by STMicroelectronics for Indoor Robotics Applications

## TL;DR

This paper evaluates the performance of the VL53L5CX ToF sensor for robotics, showing it is reliable and repeatable under various conditions.

## Contribution

The study provides a detailed characterization of the sensor's behavior under different environmental and operational conditions.

## Key findings

- The sensor shows a transient warm-up effect before stabilizing.
- Depth accuracy has a near-linear range-dependent bias with higher uncertainty for dark targets.
- Invalid measurement rates remain consistently below 10%.

## Abstract

Miniaturised proximity Time-of-Flight (ToF) sensors are attractive for robotics applications due to their low cost, compact size, and low power consumption, which makes them suitable for direct distribution on the robot body. However, both the accuracy and the reliability of their measurements are influenced by operating conditions and target properties. These aspects are not fully investigated in the manufacturer’s datasheet, yet they play a crucial role in downstream robotic tasks. To address this gap, we mounted three VL53L5CX sensors, an Ambient Light Sensor, and a thermistor on a robotic manipulator in a controlled laboratory setup and executed a series of experiments to characterise sensor performance. Specifically, experiments were conducted to quantify sensor drift over time, the influence of ambient illumination under three office lighting conditions, within-frame beam variability, depth accuracy over the 20–800 mm range for different materials, orientation sensitivity at different distances, and an empirical signal-to-noise ratio. The results reveal a transient warm-up effect at startup, after which measurements stabilise, a near-linear range-dependent bias with substantially larger uncertainty for dark targets, limited within-frame variability, and an invalid measurement rate consistently below 10%. Overall, the VL53L5CX provides repeatable measurements, and the findings of this work can be leveraged to derive more faithful sensor models, apply range bias correction, and broaden the range of robotic applications.