4WS4WD车辆的无模型自适应容错路径追踪控制
42 浏览量
更新于2024-08-26
收藏 310KB PDF 举报
本文探讨了四轮转向（4WS）和四轮驱动（4WD）道路车辆的路径跟踪控制问题，着重于开发一种无需依赖具体系统动态模型的自适应容错输出跟踪控制策略。在复杂的道路交通环境中，车辆动态变化和外部干扰常常带来挑战，这包括不确定的车辆动态以及可能发生的驱动器故障。为了克服这些问题，研究者提出了一种新颖的控制算法，它能够实时补偿这些不确定性，并能在出现意外的驱动器故障时保持有效的路径跟踪能力。
作者Dan-Yong Li、Yong-Duan Song、Dong Huang和He-Nan Chen通过理论分析和仿真验证，设计了一个能够自我适应并容忍故障的控制方案。这种方法的核心在于其对系统细节信息的需求较低，这意味着即使在缺乏精确模型的情况下也能有效运作。通过这种方式，车辆可以在面对诸如轮胎打滑、传感器失效等突发性驾驶辅助设备故障时，依然能够维持高度精确的路径追踪性能。
此外，论文还展示了将此控制策略应用到一款具有四个转向轮和四个驱动轮的实车上的实例，实验结果表明，即使在遇到转向系统故障时，该控制方法也能确保车辆稳定地沿预定路径行驶，从而提高了行驶安全性和可靠性。
"模型无关"的设计理念使得这项技术具有广泛的应用潜力，不仅适用于当前的4WS4WD车辆，也适用于不断发展的自动驾驶和智能交通系统中，对于提高系统的鲁棒性和适应性至关重要。这篇研究论文为道路车辆的故障容错控制提供了一种创新且实用的方法，有助于推动汽车行业向更高水平的安全和效率迈进。
LI et al.: FAULT-TOLERANT OUTPUT TRACKING CONTROL OF 4WS4WD ROAD VEHICLES 171
Yaw dynamics
I
z
˙
Γ=a(F
x12
sin α
f
+F
y12
cos α
f
)
− b(F
x34
sin α
r
+ F
y34
cos α
r
)
+
d
2
(−F
x12
cos α
f
+ F
y12
sin α
f
+ F
x34
cos α
r
− F
y34
sin α
r
)+f
Γ
(·) (5)
where
F
x12
=(F
x
1
+ F
x
2
) F
x34
=(F
x
3
+ F
x
4
)
F
y12
=(F
y
1
+ F
y
2
) F
y34
=(F
y
3
+ F
y
4
).
In the preceding equations, T
f
and T
r
represent the steering
torques for the front and rear wheels, respectively;
˙
θ
s
i
(i =
1−4) denote the angular velocity of each wheel; J
s
i
(i = 1−4),
J
f
, and J
r
are the moment inertia of each wheel and the
front and rear steering axles, respectively; f
s
i
(i = 1−4) are
the viscous friction coefficients of the four wheels; f
f
and f
r
denote the viscous friction coefficients of the front and rear
steering axles, respectively; k
f
and k
r
are the front and rear
steering motor torque coefficients, respectively; D is the width
of the contact surface between the wheel and the road; I
z
is the
vehicle inertia around the yaw axle; and f
θ
i
(·), f
x
(·), f
y
(·),
f
a
f
(·), and f
Γ
(·) model the external disturbing force/torque
(e.g., bearing slippage) acting on the system, which can be
assumed unknown but bounded.
From the control system point of view, the complete states
that reflect the dynamic behavior of the vehicle are the 17
variables consisting of eight rotational angle and angular ve-
locity variables of the four wheels, four lateral and longitudinal
displacement and velocity variables of the vehicle, four steering
angle and angular velocity variables of the front and rear
wheels, as well as the yaw rate variable of the vehicle, i.e.,
x =
⎡
⎢
⎢
⎢
⎣
(
˙
θ
s
1
,
˙
θ
s
2
,
˙
θ
s
3
,
˙
θ
s
4
, )
T
(θ
s
1
,θ
s
2
,θ
s
3
,θ
s
4
, )
T
(˙x
c
,x
c
, ˙y
c
,y
c
)
T
(˙α
f
,α
f
, ˙α
r
,α
r
)
T
Γ
⎤
⎥
⎥
⎥
⎦
∈ R
17
and the output vector of interest in vehicle maneuvering is y =
(
˙
θ
s
1
,
˙
θ
s
2
,
˙
θ
s
3
,
˙
θ
s
4
,α
f
,α
r
)
T
∈ R
6
. Note that the steering angles
cannot be adjusted/commanded directly from the mechanical
point of view. Instead, the adjustment of the steering angle is
accomplished through a certain motor/actuation unit. Thus, we
consider the driving torques for the wheels as the direct control
inputs, so that the following dynamic model of the vehicle can
be established without the need for small-angle approximation:
˙
x = f (x)+B(x)(u
a
+ d(x,t))
y = Cx
(6)
where u
a
=[T
1
T
2
T
3
T
4
T
f
T
r
]
T
are the actual
wheel driving/braking torques for the four wheels and the actual
steering torques for the front and rear wheels, respectively; d(·)
represents the generalized external disturbance; and f (·) is a
nonlinear function defined as
f(x)=[f
1
(·) f
2
(·) f
3
(·) ··· f
16
(·) f
17
(·)]
T
with variables explicitly defined at the bottom of the next page.
Note that any maneuvering operation of the vehicle can be
achieved by controlling the four wheels and at the same time
steering both rear and front wheels properly. Therefore, the
control objective is to generate suitable torques for the four
wheels, the two pairs of front steering wheels, and the two pairs
of rear wheels of the vehicle, so that the four wheel angular
velocities
˙
θ
s
1
,
˙
θ
s
2
,
˙
θ
s
3
, and
˙
θ
s
4
, as well as the front and rear
wheel steering angles α
f
and α
r
, are automatically adjusted to
the specific desired commands.
Remark 1: In most existing works, both α
f
and α
r
are
treated as the control inputs that can be designed/commanded
directly [18]–[26]. As such treatment leads to a non-affine
model, small-angle approximation/linearization (i.e., assuming
α small enough so that cos α = 1 and sin α = α) has to be
utilized to design the control scheme. Apparently, the perfor-
mance of the small-angle-based control scheme might not be
well maintained in practice for higher speed and larger steering
angle operations. In this paper, the steering angles, together
with the angular positions and angular velocities, are treated
as the variables to be controlled, rather than the ones able to
control, and consequently, small-steering-angle assumption is
no longer needed.
III. F
AULT-TOLERANT CONTROL DESIGN
FOR
OUTPUT TRACKING
This section focuses on designing an adaptive fault-tolerant
output tracking control scheme for a class of nonlinear systems
arisen from practical applications, including highway vehicles.
More specifically, we consider the following nonlinear system
under actuator faults and external disturbances:
˙
x = f (x)+B(x)(u
a
+ d(x,t))
y = h(x)
(7)
where x(t) ∈ R
n
, u
a
(t) ∈ R
m
, and y(t) ∈ R
r
(n ≥ m ≥ r)
are vectors of system state, control input, and output, re-
spectively; f(x) ∈ R
n
, B(x) ∈ R
n×m
, and h(x) ∈ R
r
are
nonlinear vector functions with appropriate dimensions; and
d(x,t) ∈ R
m
models external disturbances acting on the
system.
When actuator failures occur, the actual control input u
a
to
the system is no longer the designed input u. Instead, they are
related by
u
a
(t)=ρ(t)u + w(t) (8)
where ρ =diag{ρ
i
(t)} is a diagonal matrix with ρ
i
(t) ∈ (0, 1]
(i = 1, 2,...,m) being the time-varying scalar function called
the actuator efficiency factor, or the “health indicator” [19];
w(t) denotes a vector function corresponding to the portion of
the control action produced by the actuator that is completely
out of control. Obviously, such w(t) might be time varying.
Note that the case of ρ
i
= 0 corresponds to the situation that
the ith actuator fails to work completely. This case, although
out of the scope of this paper, can be addressed by using
the redundant actuation systems to reallocate the remaining
actuators to maintain the control performance.
剩余10页未读，继续阅读
2021-02-21 上传
2021-09-20 上传
2021-08-12 上传
2019-08-13 上传
2021-10-01 上传
2021-05-20 上传
weixin_38670297
- 粉丝: 7
- 资源: 927
我的内容管理
展开
最新资源
- SSM Java项目：StudentInfo 数据管理与可视化分析
- pyedgar：Python库简化EDGAR数据交互与文档下载
- Node.js环境下wfdb文件解码与实时数据处理
- phpcms v2.2企业级网站管理系统发布
- 美团饿了么优惠券推广工具-uniapp源码
- 基于红外传感器的会议室实时占用率测量系统
- DenseNet-201预训练模型：图像分类的深度学习工具箱
- Java实现和弦移调工具：Transposer-java
- phpMyFAQ 2.5.1 Beta多国语言版：技术项目源码共享平台
- Python自动化源码实现便捷自动下单功能
- Android天气预报应用：查看多城市详细天气信息
- PHPTML类：简化HTML页面创建的PHP开源工具
- Biovec在蛋白质分析中的应用：预测、结构和可视化
- EfficientNet-b0深度学习工具箱模型在MATLAB中的应用
- 2024年河北省技能大赛数字化设计开发样题解析
- 笔记本USB加湿器：便携式设计解决方案
