Hybrid Dynamics

Homework 2

Josef Nasim 208667139

David Dayan 941170037

May 31, 2021

1
Task 1

We are asked to produce a function that will produce the minimum and maximum position along the x-axis
and produce a plot showing the region where the force of gravity can be placed in which the system will
remain static.
The function is function [xmin , xmax ] = friceq (r, α, mu)

Where,
45◦
     
−1 2 0.4
r= α= ◦ mu =
1 2 135 0.4

We will solve this problem as recommended using the linprog function given in matlab.

Where linprog is [X,FVAL] = linprog(c,A,b,Aeq,beq)

We need to find all the inputs in the function to find the range of values that will work.

But first let us find and define a couple terms.

1. The matrix containing

 the vector directions
 of the normal to each plane(where the first column vector
cos(45) cos(135)
corresponds tor1 n=sin(45) sin(135)
0 0
 
0.7071 0.7071
2. The matrix containing the vector directions along the direction of the plane. t=−0.7071 0.7071
0 0

3. .The contact forces for each plane will be comprised of vector f = [fix , fiy ]

4. The 2d inequlaities for a friction cone for each plane is defined as −µf · n ≤ f · t ≤ µf · n

A and b will be the system that contains all the inequalities of the system such that Ax ≤ b .We will get
the inequalities from the inequalities of 2D friction cone that were defined.

We will now define the vector x that will contain all the componenents we wish to use in this problem
either in equalities(later) or the inequalities. x = [f1x , f1y , f2x , f2y xc ]T where xc is the positions where the
center of mass can.From here we recieve that:
 
  f1x  
(t1 − µn1 ) · e1 (t1 − µn1 ) · e2 0 0 0   0
f1y   
 0 0 (t2 − µn2 ) · e1 (t2 − µn2 ) · e2 0   0
 

(−t1 − µn1 ) · e1 (−t1 − µn1 ) · e2 f2x  ≤  
0 0 0 f2y  0
0 0 (−t2 − µn2 ) · e1 (−t2 − µn2 ) · e2 0 0
xc

2
 Next we will find Aeq and beq using the equality Aeq x = beq These we will get by formulating the equations
of motion for equilibrium.
Which are:
f1 + f2 − mge2 = 0

e3 · [r1 × f1 + r2 × f2 + rc × (−mge2 )] = 0
Putting this into marix form, we get that:


  f 1x  
1 0 1 0 0 f1y 
  0
 0 1 0 1 0 f2x  = mg
    
−r(2, 1) r(1, 1) −r(2, 2) r(1, 2) −mg f2y  0
xc

We then took mg=1. As all the vectors and matrices must be numerical to perform the function.

Lastly we choose cmin = [0, 0, 0, 0, 1]T . As we choose to find the minimum position along the x-axis that
satisfy all the equalities and inequalities will hold. As we wish to minimize the function cT x
For the maximum we will choose cmax = [0, 0, 0, 0, −1]T . As we know that the maximum of the function at
a point is the same as the function at the negative of that point.
Using the linprog function we find the permissible range for which the COM can be placed along the X
axis.
The corresponding diagrams for each value of µ are :

For µ = 0.4

3
For µ = 0.8

Task 2

Use the moment labeling method in order to graphically find the permissible region of COM position for
the example in Figure 2 in both cases (a)+(b), and compare to the MATLAB function’s output. Use either
handwriting or any graphical software.

Below we have the graphs for the moment labelling method where α and β are the half angles from
their respective normal. The value of α and β is given by α = β = arctan(µ)

For µ = 0.4 =⇒α = β = 21.8◦

4
For µ = 0.8 =⇒α = β = 38.66◦

5
As we can see our results from the moment labelling method are basically identical to our results from
the linear programming method used for task 1. This is a very insightful result as it means problems such
as this do not necessarily require heavy math but only a simple graphical analysis.

Task 3

For the example of figure 2 and equal friction coefficient µ1 = µ2 = µ find the range of µ for which the
COM equilibrium region is a semi-infinite half-plane (bounded only from one side), and the range of µ for
which the COM equilibrium region covers the entire plane. Explain the results graphically and validate by
output of the computation in task 1.

To find the ranges of µ where the COM can be placed in a semi-infinite half-plane and infinite plane
we took a graphical approach to solve the task. We began by drawing a couple of moment labelling for
different angles α which is the half-angle from the normal directions which shares the relation α = arctan(µ).

6
From this initial drawings we see that as α grows, P − gets smaller and its also easy to see that once both
inner friction lines are parallel P − will be the empty set.

7
Once we drew the moment labelling for α = 45◦ we saw that indeed P − is the empty set which means that
the COM can be placed anywhere to the right of P + making it the lower bound for µ for which the COM
can be place on a semi-infinite half-plane. To find the extent of this range we drew more drawings with
higher values of α.

From the drawings we can see that as α grows the area of P + gets smaller and its easy to see that once
the inner friction line of r2 intersects with r1 , then P + and P − will both be part of the empty set, once we

8
exceed that angle the COM can be placed anywhere in the infinite plane.

In conclusion from the moment labelling method drawings we can create the ranges of µ for which the
COM can be placed on the semi-infinite half-plane and infinite plane.

(
45◦ ≤ α < 63.43◦ semi − inf inite
63.43◦ < α inf inite
( =⇒α = arctan(µ)
1 ≤ µ < 2 semi − inf inite
2<µ inf inite

This results were also compared numerically and we achieved the exact results.

Task 4

In the same fashion we did in task 1. We will do another optimization. the difference being is that our
vector x will now be of the form x = [f1x , f1y , f2x , f2y Fx ]T .Where FX is horizontal force in which we want
to find the minimum and maximum range for the system to remain in equilibirum.
From this fact we have to change only our equalities as the Fx has no effect on the inequalities due to the
friction jsut for our equations of motion, so A and b will remian unchanged.

Our equations of motion in matrix form of Aeq x = beq is:



  f 1x  
1 0 1 0 1 f1y 
  0
 0 1 0 1 0 f2x  = mg
    
−r(2, 1) r(1, 1) −r(2, 2) r(1, 2) −5mg f2y  mg
Fx

This equations take into force produced by the mass of the object by the COM.

Taking into account that mg =1. We can use the function linprog to find the range of values of Fx
that will satisfy the equalities and inequalities, using the same vectors of cmin and cmax used previously.
The values we achieve is that
Our results from the computational method return that the range where Fx is permissible is

−0.500 < Fx < 0.325[N ]

Our diagram for the graphical method is shown below

9
To understand the permissible values of Fx we use the the angles that intersects with P + and P − with the
relationship of tan(θ) = FFxy = −mg
Fx
. from this we can get the permissible range for Fx

=⇒ −0.500 < Fx < 0.325[N ]

And as we can see they are the same for both methods.

Task 5

For the example of figure 2 and equal friction coefficient µ1 = µ2 = 0.4 it is now given that the COM is
located at (1,h) where 0 10. In addition to gravity force of 1N, a disturbance force fd having unknown
direction acts at the COM. For h=5, find the maximal value of the magnitude fd for which a static equi-
librium solution exists under any direction of fd. Explain. Using MATLAB computation, plot a graph of
the maximal magnitude fd as a function of h. (No need to derive an analytic expression. Compute for a
discrete series of values of h.)

For the first task we used the Linear Programming function in order to find the maximal magnitude of the
force Fd for every direction taking intervals of 2[degrees]. The matrices we used for the linprog function are
as follows:
  
(t1 − µn1 ) · e1 (t1 − µn1 ) · e2 0 0 0 f1x  
0

 0 0 (t2 − µn2 ) · e1 (t2 − µn2 ) · e2 0  f1y  0
 
(−t1 − µn1 ) · e1 (−t1 − µn1 ) · e2 0 0 0  f2x  ≤ 0
   

 0 0 (−t2 − µn2 ) · e1 (−t2 − µn2 ) · e2 0  f2y 
0
0 0 0 0 −1 Fd

10
 
 f1x  
1 0 1 0 cos θ f1y 
  0
 0 1 0 1 sin θ  f2x  = mg  cmax = [0, 0, 0, 0, −1]T
 
−r(2, 1) r(1, 1) −r(2, 2) r(1, 2) sin(θ) − h cos(θ) f2y  mg
Fd

Where θ is the angle of the direction of the force with respect to e1 . From the Linear programming we
obtain the following graph which shoes the maximal magnitude of Fd for each direction given by θ.

The maximal value of the magnitude of Fd will be the minimal of this graph as it will be accepted for
any direction. The value we found for this case when h = 5, kFd k ≤ 0.3091[N ] will have a solution for the
statics problem always.

To find the maximal magnitude for every h, 0 ≤ h ≤ 10 we simply ran numerical computations similar
to the previous question with intervals of 0.1 for h and obtained the following graph.

11
Summary

12