A Haptic-Based Proximity Sensing System for Buried Object in Granular Material

Zeqing Zhang,Ruixing Jia,Youcan Yan,Ruihua Han,Shijie Lin,Qian Jiang,Liangjun Zhang,Jia Pan

2024-11-26

Abstract:The proximity perception of objects in granular materials is significant, especially for applications like minesweeping. However, due to particles' opacity and complex properties, existing proximity sensors suffer from high costs from sophisticated hardware and high user-cost from unintuitive results. In this paper, we propose a simple yet effective proximity sensing system for underground stuff based on the haptic feedback of the sensor-granules interaction. We study and employ the unique characteristic of particles -- failure wedge zone, and combine the machine learning method -- Gaussian process regression, to identify the force signal changes induced by the proximity of objects, so as to achieve near-field perception. Furthermore, we design a novel trajectory to control the probe searching in granules for a wide range of perception. Also, our proximity sensing system can adaptively determine optimal parameters for robustness operation in different particles. Experiments demonstrate our system can perceive underground objects over 0.5 to 7 cm in advance among various materials.

Robotics,Fluid Dynamics,Geophysics,Instrumentation and Detectors

What problem does this paper attempt to address?

### What problems does this paper attempt to solve?

This paper aims to address the challenges of sensing buried objects in granular materials (such as sand, soil, etc.). Specifically, existing proximity sensors face the following problems when dealing with buried objects in granular materials:

1. **High cost**: High costs due to complex hardware.
2. **Unintuitive results**: It is difficult for users to understand the information output by the sensors, increasing the cost of use.
3. **Limited signal propagation**: Unlike air and liquid, the opacity and complex physical properties of granular materials greatly reduce the effectiveness of traditional sensors (such as ultrasonic or infrared).

To solve these problems, the author proposes a new proximity sensing system based on tactile feedback - GRAINS (Granular - Material - embedded Autonomous Proximity Sensing System) for early sensing of the presence of buried objects in granular materials. The system utilizes the unique "failure wedge zone" phenomenon in granular materials and combines Gaussian Process Regression (GPR) to identify changes in force signals caused by object approach, thereby achieving near - field sensing.

### Main contributions

1. **Proposed an invasive proximity sensing system**: GRAINS learns based on the real - time interactive force signals between the probe and the granules.
2. **Discovered the role of the failure wedge zone**: This zone can act as an "airbag layer" for the probe, preventing direct collisions between the probe and the buried object.
3. **Designed the trajectory for controlling probe movement**: By designing a spiral trajectory, the sensing range is expanded, and the tactile signals between the failure wedge zone and underground objects are studied.
4. **Achieved autonomous parameter calibration**: The system can adaptively update the optimal parameters to ensure robust operation in different types of granular materials.

### Mathematical formula representation

The mathematical formulas involved in the paper are as follows:

- **Measurement set**:
\[
T=\{(x_t, f_t)|t = 1,\cdots,L\}
\]
where \(x_t\in W_g\) is the position of the sensor, and \(f_t = [f_{tx}, f_{ty}]\) is the reading of the dynamometer on the \(x\) and \(y\) axes.

- **Near - field sensing function**:
\[
\xi(T)=
\begin{cases}
1 & \text{when an object is approaching}\\
0 & \text{otherwise}
\end{cases}
\]

- **Proximity sensing range**:
\[
\zeta(\xi)=\min||x_o - x_L||_2
\]
where \(x_o\in O_i\) represents the position of the buried object, and \(x_L\in T|t = L\) represents the final position of the sensor.

- **Force pattern definition**:
\[
F=\{(t, f_t)|f_t=\sqrt{(f_{tx})^2+(f_{ty})^2},t = 1,\cdots,M\}
\]

- **Gaussian process regression model**:
\[
p(t)\sim\mathcal{GP}(m(t)\equiv0, k(t, t'))
\]
\[
\epsilon_t\sim\mathcal{N}(0,\sigma^2)
\]
\[
f = p(t)+\epsilon
\]

- **Predict the force value distribution of new inputs**:
\[
p(t^*|F)\sim\mathcal{GP}\left(K(t^*, t)[K(t, t)+\sigma^2I]^{-1}f, K(t^*, t^*) - K(t^*, t)[K(t, t)+\sigma^2I]^{-1}K(t, t^*)\right)