Editorial Type:
Article Category: Research Article
 | 
Online Publication Date: 01 Oct 2016

Estimation of the Tire Contact Patch Length and Normal Load Using Intelligent Tires and Its Application in Small Ground Robot to Estimate the Tire-Road Friction

,
, and
Page Range: 248 – 261
DOI: 10.2346/tire.16.440402
Save
Download PDF

ABSTRACT

Tire-road friction estimation is one of the most popular problems for the tire and vehicle industry. Accurate estimation of the tire-road friction leads to better performance of the traction and antilock braking system controllers, which reduces the number of accidents. Several researchers have worked in the field of friction estimation, and many tire models have been developed to predict the tire-road friction. In this article, an intelligent tire, which has an embedded accelerometer placed on the inner liner of the tire, is used to estimate the tire contact patch length parameter and normal load. To accomplish this, first, an existing tire testing trailer equipped with a force hub to measure tire forces and moments, a high-accuracy encoder to measure the angular velocity of the wheel, and VBOX, which is a global positioning system–based device, to estimate the longitudinal speed of the trailer was used. As a practical application for the normal load algorithm, a wheeled ground robot, which is equipped with several sensors, including an accelerometer and a flexible strain sensor inside the tire (used for terrain identification purposes), was designed and built. A set of algorithms was developed and used with the test data that were collected with both the trailer and the robot, and the contact patch length and the normal load were estimated. Also, the friction potential between the tire and the road was evaluated using a small ground robot.

Keywords: intelligent tire; contact patch; normal load; friction; force estimation; neural network
FIG. 1
FIG. 1

Trailer setup that is towed by a truck.

FIG. 2
FIG. 2

The quarter car setup and data-collecting and control system.

FIG. 3
FIG. 3

The accelerometer inside the tire.

FIG. 4
FIG. 4

The schematic of the robot: 1: DC motor, 2: USB- NI DAQ, 3: laptop, 4: signal conditioner for the accelerometers, 5: motor controller, 6: current sensor, 7: single axis accelerometer, 8: slip ring, 9: triaxial accelerometer, 10: rotary encoder, 11: 0.248 m pneumatic tire, 12: mechanism to hold the encoder, 13: 24-V battery.

FIG. 5
FIG. 5

The schematic of robot's data-collecting system.

FIG. 6
FIG. 6

Radial and circumferential component of acceleration.

FIG. 7
FIG. 7

Normal force estimation algorithm.

FIG. 8
FIG. 8

Normal force, longitudinal speed, and encoder signal (raw data).

FIG. 9
FIG. 9

Two-layer feed-forward network, used to fit measured data.

FIG. 10
FIG. 10

Free body diagram of single wheel model in acceleration mode [15].

FIG. 11
FIG. 11

Left: Histogram of error percentages for the training data set. Right: Histogram of error percentages for the validation data set.

FIG. 12
FIG. 12

Measured and estimated wheel speed.

FIG. 13
FIG. 13

Measured and estimated longitudinal robot speed.

FIG. 14
FIG. 14

The friction coefficient for different slip.

Contributor Notes

1 Corresponding author. Department of Biomedical Engineering and Mechanics, Virginia Tech, Blacksburg, Virginia 24061, USA. Email: meysam@vt.edu
  • Download PDF