State-observers for semi-active suspension control applications with low sensitivity to unknown road surfaces

Diana Hernandez-Alcantara, Luis Amezquita-Brooks, Ruben Morales-Menendez

Research output: Contribution to conferencePaper

Abstract

Semi-active suspension systems aim to improve the vehicle safety and comfort. For these systems control laws are required to achieve the desired performance improvements. On the other hand, the instrumentation of the vehicle suspension typically consists only in accelerometers, which are used to measure the vertical accelerations. However, velocities and/or displacements are required to implement the most common control algorithms for semi-active suspension systems. For instance, Skyhook and Groundhook controllers require the knowledge of the suspension vertical velocities. In this article several vertical velocities estimation approaches are studied and compared. In practical applications, it is common to use simple integrators to estimate these variables; nonetheless, it is well known that integrator-based estimations present errors due to drift. In applications where high performance is required, a better estimation of the state variables of the suspension system is essential. An additional problem for the estimation process is that the system is also affected by an unknown input: the road surface. This complicates the use of traditional observer schemes, thus an Unknown Input Observer (UIO) can be considered. There are many theoretical reports dealing with state estimation in presence of unknown inputs; however, only a few of them apply these techniques to vehicle suspension systems. An observer capable of estimating the unmeasured state variables of the Quarter of Vehicle (QoV) dynamics subject to unknown road surfaces is proposed. It is shown that the decoupling of the unknown input induces observability problems, which can be overcome by reducing the observer performance. A frequency-domain analysis shows that the resulting observer presents a good level of disturbance rejection of the unknown road. In addition, it is shown that the use of frequency-analysis tools can help to elucidate how to reduce the effect of the road surface even for classical observer schemes. Early results show good performance of the proposed observer in several scenarios. A comparison between several classical estimation methods is also included. Copyright © 2014 SAE International.
Original languageEnglish
DOIs
Publication statusPublished - 1 Jan 2014
Externally publishedYes
EventSAE Technical Papers -
Duration: 1 Jan 2014 → …

Conference

ConferenceSAE Technical Papers
Period1/1/14 → …

Fingerprint

Active suspension systems
Vehicle suspensions
Frequency domain analysis
Disturbance rejection
Observability
State estimation
Accelerometers
Error analysis
Control systems
Controllers

Cite this

Hernandez-Alcantara, Diana ; Amezquita-Brooks, Luis ; Morales-Menendez, Ruben. / State-observers for semi-active suspension control applications with low sensitivity to unknown road surfaces. Paper presented at SAE Technical Papers, .
@conference{962f920fb6264a12b1f969a10655f332,
title = "State-observers for semi-active suspension control applications with low sensitivity to unknown road surfaces",
abstract = "Semi-active suspension systems aim to improve the vehicle safety and comfort. For these systems control laws are required to achieve the desired performance improvements. On the other hand, the instrumentation of the vehicle suspension typically consists only in accelerometers, which are used to measure the vertical accelerations. However, velocities and/or displacements are required to implement the most common control algorithms for semi-active suspension systems. For instance, Skyhook and Groundhook controllers require the knowledge of the suspension vertical velocities. In this article several vertical velocities estimation approaches are studied and compared. In practical applications, it is common to use simple integrators to estimate these variables; nonetheless, it is well known that integrator-based estimations present errors due to drift. In applications where high performance is required, a better estimation of the state variables of the suspension system is essential. An additional problem for the estimation process is that the system is also affected by an unknown input: the road surface. This complicates the use of traditional observer schemes, thus an Unknown Input Observer (UIO) can be considered. There are many theoretical reports dealing with state estimation in presence of unknown inputs; however, only a few of them apply these techniques to vehicle suspension systems. An observer capable of estimating the unmeasured state variables of the Quarter of Vehicle (QoV) dynamics subject to unknown road surfaces is proposed. It is shown that the decoupling of the unknown input induces observability problems, which can be overcome by reducing the observer performance. A frequency-domain analysis shows that the resulting observer presents a good level of disturbance rejection of the unknown road. In addition, it is shown that the use of frequency-analysis tools can help to elucidate how to reduce the effect of the road surface even for classical observer schemes. Early results show good performance of the proposed observer in several scenarios. A comparison between several classical estimation methods is also included. Copyright {\circledC} 2014 SAE International.",
author = "Diana Hernandez-Alcantara and Luis Amezquita-Brooks and Ruben Morales-Menendez",
year = "2014",
month = "1",
day = "1",
doi = "10.4271/2014-01-0867",
language = "English",
note = "SAE Technical Papers ; Conference date: 01-01-2014",

}

State-observers for semi-active suspension control applications with low sensitivity to unknown road surfaces. / Hernandez-Alcantara, Diana; Amezquita-Brooks, Luis; Morales-Menendez, Ruben.

2014. Paper presented at SAE Technical Papers, .

Research output: Contribution to conferencePaper

TY - CONF

T1 - State-observers for semi-active suspension control applications with low sensitivity to unknown road surfaces

AU - Hernandez-Alcantara, Diana

AU - Amezquita-Brooks, Luis

AU - Morales-Menendez, Ruben

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Semi-active suspension systems aim to improve the vehicle safety and comfort. For these systems control laws are required to achieve the desired performance improvements. On the other hand, the instrumentation of the vehicle suspension typically consists only in accelerometers, which are used to measure the vertical accelerations. However, velocities and/or displacements are required to implement the most common control algorithms for semi-active suspension systems. For instance, Skyhook and Groundhook controllers require the knowledge of the suspension vertical velocities. In this article several vertical velocities estimation approaches are studied and compared. In practical applications, it is common to use simple integrators to estimate these variables; nonetheless, it is well known that integrator-based estimations present errors due to drift. In applications where high performance is required, a better estimation of the state variables of the suspension system is essential. An additional problem for the estimation process is that the system is also affected by an unknown input: the road surface. This complicates the use of traditional observer schemes, thus an Unknown Input Observer (UIO) can be considered. There are many theoretical reports dealing with state estimation in presence of unknown inputs; however, only a few of them apply these techniques to vehicle suspension systems. An observer capable of estimating the unmeasured state variables of the Quarter of Vehicle (QoV) dynamics subject to unknown road surfaces is proposed. It is shown that the decoupling of the unknown input induces observability problems, which can be overcome by reducing the observer performance. A frequency-domain analysis shows that the resulting observer presents a good level of disturbance rejection of the unknown road. In addition, it is shown that the use of frequency-analysis tools can help to elucidate how to reduce the effect of the road surface even for classical observer schemes. Early results show good performance of the proposed observer in several scenarios. A comparison between several classical estimation methods is also included. Copyright © 2014 SAE International.

AB - Semi-active suspension systems aim to improve the vehicle safety and comfort. For these systems control laws are required to achieve the desired performance improvements. On the other hand, the instrumentation of the vehicle suspension typically consists only in accelerometers, which are used to measure the vertical accelerations. However, velocities and/or displacements are required to implement the most common control algorithms for semi-active suspension systems. For instance, Skyhook and Groundhook controllers require the knowledge of the suspension vertical velocities. In this article several vertical velocities estimation approaches are studied and compared. In practical applications, it is common to use simple integrators to estimate these variables; nonetheless, it is well known that integrator-based estimations present errors due to drift. In applications where high performance is required, a better estimation of the state variables of the suspension system is essential. An additional problem for the estimation process is that the system is also affected by an unknown input: the road surface. This complicates the use of traditional observer schemes, thus an Unknown Input Observer (UIO) can be considered. There are many theoretical reports dealing with state estimation in presence of unknown inputs; however, only a few of them apply these techniques to vehicle suspension systems. An observer capable of estimating the unmeasured state variables of the Quarter of Vehicle (QoV) dynamics subject to unknown road surfaces is proposed. It is shown that the decoupling of the unknown input induces observability problems, which can be overcome by reducing the observer performance. A frequency-domain analysis shows that the resulting observer presents a good level of disturbance rejection of the unknown road. In addition, it is shown that the use of frequency-analysis tools can help to elucidate how to reduce the effect of the road surface even for classical observer schemes. Early results show good performance of the proposed observer in several scenarios. A comparison between several classical estimation methods is also included. Copyright © 2014 SAE International.

U2 - 10.4271/2014-01-0867

DO - 10.4271/2014-01-0867

M3 - Paper

ER -