1. The document models a self-balancing two-wheeled vehicle as an inverted pendulum with a moving base. It presents both nonlinear and linearized models.
2. The linearized model approximates the vehicle balancing around zero degrees and replaces various forces with a single friction force proportional to velocity. This allows deriving state space equations and transfer functions.
3. The state space representation models the vehicle's position, velocity, angle, and angular velocity. PID and LQR controllers are then briefly mentioned to control the linearized system.
1. The document models a self-balancing two-wheeled vehicle as an inverted pendulum with a moving base. It presents both nonlinear and linearized models.
2. The linearized model approximates the vehicle balancing around zero degrees and replaces various forces with a single friction force proportional to velocity. This allows deriving state space equations and transfer functions.
3. The state space representation models the vehicle's position, velocity, angle, and angular velocity. PID and LQR controllers are then briefly mentioned to control the linearized system.
1. The document models a self-balancing two-wheeled vehicle as an inverted pendulum with a moving base. It presents both nonlinear and linearized models.
2. The linearized model approximates the vehicle balancing around zero degrees and replaces various forces with a single friction force proportional to velocity. This allows deriving state space equations and transfer functions.
3. The state space representation models the vehicle's position, velocity, angle, and angular velocity. PID and LQR controllers are then briefly mentioned to control the linearized system.
1. The document models a self-balancing two-wheeled vehicle as an inverted pendulum with a moving base. It presents both nonlinear and linearized models.
2. The linearized model approximates the vehicle balancing around zero degrees and replaces various forces with a single friction force proportional to velocity. This allows deriving state space equations and transfer functions.
3. The state space representation models the vehicle's position, velocity, angle, and angular velocity. PID and LQR controllers are then briefly mentioned to control the linearized system.
The two wheeled self balancing vehicle is modelled as an inverted pendulum with a moving bottom.
1.1
Nonlinear model
In the nonlinear model air drag and rolling resistance are taken in account. And another external force is also applied which should simulate a person standing and pushing forward on the vehicle. A Figure of the model which is used, see Figure 1.
Figure 1: Model of the Segway with air resistance FD , rolling resistance Fr and the force (a moment) applied by a person on the rod FB . List of all variables g := Gravity m1 := Mass of wheels m2 := Mass of pendulum Iw := Moment of inertia of the wheels Ip := Moment of inertia of the pendulum r := Radius of wheels l := Distance to center of gravity for the pendulum kD := Constant for air resistance b := Constant for rolling resistance F := Force applied by the vehicle FD := Air resistance Fr := Rolling resistance FB := Body force N := Normal force
The air- and rolling resistance has been modelled as followed
Nb r = sign(x)k D x
Fr = sign(x)
FD
where sign is a function which gives the sign of a number, in this case the velocity. This is used because the air-and rolling resistance are applied in the opposite direction in which the vehicle is moving. The Lagrangain L = T V of the system is given by L=
which lead to the final equations of motions for the nonlinear model x
+ (m1 + m2 ) x + m2 l cos() m2 l2 sin() = F FD Fr r2 Ip + m2 l x cos() + m2 l2 m2 glsin() = FB l FD l cos().
Iw
1.2
(1) (2)
Linear model
The main goal of using a linear model instead of the nonlinear is to find the state space of the system therefore is now the air drag and rolling resistance replaced by an new friction force Ff which is proportional to the velocity and therefore is always acting in the right direction. This is because before the function sign was used to determine the direction of both air drag and rolling resistance but since there is now easy way of determine the Laplace of the function sign this change was done. The before applied body force is now removed but is later going to be added as an disturbance. To clarify the new model an the acting forces, see Figure 2.
Figure 2: Model of the Segway with a general friction force Ff applied.
The friction force is modelled as followed Ff = kf x kf := Constant for friction. The Lagrangain L = T V of the system is still given by L=
+ (m1 + m2 ) x + m2 l cos() m2 l2 sin() = F Ff r2 Ip + m2 l x cos() + m2 l2 m2 glsin() = 0.
As one can see there are still nonlinear terms in the above equations. To get rid of these the assumption is made that the angle is small since the goal is to balance the vehicle around zero degrees. Which means that following approximation can be made cos() = 1 sin() = 2 0.
This leads then to the final linear equations of motions
Iw
+ (m1 + m2 ) x + m2 l = u kf x r2
(3)
Ip + m2 l x + m2 l2 m2 gl = 0.
(4)
Note that F (the input) has been replace with u and Ff with kf x.
1.3
Transfer function
In order to find the transfer functions for linear system it is first necessary to Laplace transform 3 and 4. After that the equations can be rewritten so that the transfer functions can be found.
1.4
State Space
To rewrite the linear system on state space form it is first necessary to rewrite 3 and 4 so that the two equations only contain one variable and there respective derivatives. Which with some algebra leads to these equations x =
=. The linear system is represented in state space form as following
