Contents

1 Topological concept
1.1 Topology Concepts . . . . .
1.1.1 Fundamental . . . .
1.1.2 Algebralic topology .
1.1.3 Fun stuffs . . . . . .

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

2
2
2
3
4

. . . . .
. . . . .
. . . . .
problem

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

5
5
5
6
6

3 Geometric representation
3.1 Geometric modelling . . . . . . . . . . . . . . . . . . . . . . . . .

7
7

4 Math, mostly algebra

4.1 Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2 Polynomial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3 Varieties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

8
8
8
9

2 Configuration Space
2.1 2D . . . . . . . . . .
2.2 3D . . . . . . . . . .
2.3 Obstacle space . . .
2.4 The motion planning

Chapter 1

Topological concept
1.1
1.1.1

Topology Concepts
Fundamental

A set X is called a topological space if there is a collection of subsets of X

called open sets for which the following axioms hold:
The union of any number of open sets is an open set.
The intersection of a finite number of open sets is an open set.
Both X and are open sets.
Thus,the complement of any open set is closed, and the complement of any
closed set is open. Any closed interval, such as [0, 1], is a closed set because
its complement, (, 0)U (1, ), is an open set.
Hausdorff axiom: For any distinct x1, x2belongX, there exist open sets
O1 and O2 such that x1belongO1 , x2belongO2 , and O1 and O2 are separate.
Continuous functions: A very simple definition of continuity exists for
topological spaces. Function f is called continuous if f 1 (O) is an open set
belong to X for every open set O belong to Y
Homeomorphism: Donut equals cup. Between space X and Y, if there is
a function f so that both f and f 1 are continuous, then X and Y are homeomorphic.
Manifold:Manifold M that belong to Rm . For every x belong to M, there
exists open set O, so that x belong to O and O is homeomorphic to Rn . Another
explaination: for every x belong to M, in the vicinity of x, the space is just like
Rn . Some manifold:
R: (0,1)
S: a circle
R2 : (0,1)x(0,1) : a square
2

 SxS : a torus
S 2 : a sphere
RxS : a flat cylinder
R/ : = S : connect the two heads of a straight line, we have a circle.
Idenfitifcation.
R2 / : depends on the identification, we can be Mobius band, sphere,
Torus, RP 2
n-dimensional projective space is very important. RP n is the set of all
lines in Rn+1 that passes through the origin. each line is considered as a point in
RP n . If we consider the n-dimensional sphere S n , then each line will intersect
with the sphere at two points. If this two points are identified, the sphere will
be homeomorphic to RP n .
Paths Let X be a topological space, which for our purposes will also be a
manifold. A path is a continuous function, : [0, 1]maptoX.
Path connected A topological space X is said to be path connected if for
every two points x1 x2 belong to X, there exists a path connecting them.
Homotopic Paths Two paths with the same end points are said to be
homotopic if they can be morphed to become the other one continuously. This
mean if there is a hole in a R2 manifold, two paths that are separate by the
hold are not homotopic.
equivalent path for x1 and x2 both belongs to X, if the paths are all
homotopic, they are said to by simply connected. And if for all point that
belongs to X, the they are all simply connected, then X is considered simply
connected. If it is not the case, then X is said to be multipy connected.

1.1.2

Algebralic topology

Groups Algebraic topology, which is a branch of mathematics that characterizes the structure of topological spaces in terms of algebraic objects, such as
groups. These resulting groups have important implications for motion planning. Therefore, we give a brief overview. First, the notion of a group must be
precisely defined. A group is a set, G, together with a binary operation x such
that the following group axiom are obeyed:
Closure for any a and b belong to G, a x b also belongs to G
Associativity a x b x c = (a x b) x c = a x (b x c)
Identity i x a = a, and a x i = a
Inverse axa1 = i
The fundamental group Now an interesting group will be constructed from
the space of paths and the equivalence relation obtained by homotopy. The

funda- mental group, 1 (X) (or first homotopy group), is associated with any
topological space, X. Let a (continuous) path for which f (0) = f (1) be called
a loop. Let some xb belongs to X be designated as a base point. For some
arbitrary but fixed base point, xb , consider the set of all loops such that f(0)
= f(1) = xb . This can be made into a group by defining the following binary
operation. Let 1 : [0, 1] - X and 2 : [0, 1] - X be two loop paths with the
same base point. Their product is defined as:
= 1 (2t)(t < 1/2)

(1.1)

= 2 (2t 1)(t >= 1/2)

(1.2)

For a simply connected space, there is only one class and therefore the fundamental group consist of only the identify element.
For X = S, there is a class for each integer i belong to Z. i0 means the path
class winds i time around S in the counterclockwise direction, while i0 means
the path class winds in the clockwise direction. Thus 1 (S 1 ) = Z
in a similar fashion, the fundamental group of T n is Z n . For each i corresponding to the n-the dimension of the torus, the loop goes around the n-th circle i
times.
The fundamental Group of RP 2 1 (RP 2 ) = Z2 , which stands for the cyclic
group of order 2, which correspond to addition mod 2 (operation of the group).
There are only two equivalent classes, and they are identified based on the number of loops being odd or even. Higher order homotopy Group, instead of a
line in the case of 1st order homotopy group, 2nd order homotopy group generate a sheet/surface connecting two points. And therefore, it can detect bubble
in R3 space.

1.1.3

Fun stuffs

For a car that is driving on a big sphere (like the earth) C space is actually
SO(3) not S 2 S

Chapter 2

Configuration Space
2.1

2D

For a rigid body in 2D plane, translation yield a manifold of R2 while rotation

yield a manifold of S. Therefore C = R2 S 1
Matrix group The set of all nonsingular n x n real-valued matrices is called the
general linear group, denoted by GL(n) (matrix multiplication, identify matrix
and inverse matrix).
Orthogonal group O(n) is the set of orthogonal matrices: AAT = I. These
matrices have orthogonal columns (inner product of each pair is zero) and determinant is always 1 or -1.
Special Orthogonal group SO(n) is the subgroup of O(n) in which every
matrix has determinant 1. - the group of n-dimensional rotation matrices.
2
The set of all n x n matrices belong to Rn because n2 entries in the matrix
can be independently chosen. For GL(n), singular matrices are removed but it
2
is still homeomorphic to Rn .
For O(n), because AAT = I, there are n2 unknowns with n2 equations to solve.
However, there are redundant equations for each column vector should only be
multiplied with each row vector only once. There are only n(n+1)/2 equations
and therefore the dimension is n(n-1)/2. For SO(n), there are only half of the
matrices of O(n), it still has n(n-1)/2 dimensions.
SO(2) SO(2) is the manifold of rotation in the 2D space. SO(2) is homeomorphic to S. (rotation about an axis). O(2) is actuall homeomorphic to 2 times S
(two circles)
Complex numbers can be used to represent rotation. Complex number a + bi
can be used to represent SO(2).

2.2

3D

In this section, rotation group SO(3) is frequently mentioned. It is homeomorphic to half-sphere S 3 . We could identify two antinodal points to achieve SO(3)
5

with S 3 . The manifold SO(3) belongs is said to be homeomorphic to RP 3 ,

which is the set of all lines going through the origin of the 4D space R4 .
quarternion is a convenient way to represent SO(3). Similar to how complex
number can represent a circle, quarternion represents an S 3 circle completely.

2.3

Obstacle space

The obstacle region, Cobs C, is defined as:

Cobs = {q C | A(q) O =
6 }

(2.1)

Obstacle space in 2D and 3D space Calculated using Minowsky different.

X Y = {x y Rn | x X

and

y Y}

(2.2)

In terms of the Minkowski difference, Cobs = O A(0)

2.4

The motion planning problem

Formulation 4..1 (The Piano Movers Problem)

A world W in which either W = R2 or W = R3 .
A semi-algebraic obstacle region O W in the world.
A semi-algebraic robot is defined in W. It may be a rigid robot A or a
collection of m links, A1 , A2 , . . . , Am .
The configuration space C determined by specifying the set of all possible
transformations that may be applied to the robot. From this, Cobs and
Cf ree are derived.
A configuration, qI Cf ree designated as the initial configuration.
A configuration qG Cf ree designated as the goal configuration. The
initial and goal configurations together are often called a query pair (or
query) and designated as (qI , qG ).
A complete algorithm must compute a (continuous) path, : [0, 1] Cf ree ,
such that (0) = qI and (1) = qG , or correctly report that such a path does
not exist.
It was shown by Reif [817] that this problem is PSPACE-hard, which implies
NP-hard. The main problem is that the dimension of C is unbounded.

Chapter 3

Geometric representation
3.1

Geometric modelling

Polygonal model in 2D space, a convex Polygonal model if and only if, for
any pair of points in X, all points along the line segment that connects them
are contained in X.
Polyhedral model similar to Polygonal model but in 3D space
Semi-algebralic model Instead of straight line, we can also use continous line
to define a model. Basically we use function f to describe the boundary.
H = {(x, y) W | f (x, y) 0}.

Chapter 4

Math, mostly algebra

4.1

Field

Keep in mind several familiar examples of fields: the rationals, Q; the reals, R;
and the complex plane, C. You may verify that these fields satisfy the following
six axioms.
A field is a set F that has two binary operations, : F F F (called
multiplication) and + : F F F (called addition), for which the following
axioms are satisfied:
(Associativity) For all a, b, c F, (a + b) + c = a + (b + c) and (a b) c =
a (b c).
(Commutativity) For all a, b F, a + b = b + a and a b = b a.
(Distributivity) For all a, b, c F, a (b + c) = a b + a c.
(Identities) There exist 0, 1 F, such that a + 0 = a 1 = a for all a F.
(Additive Inverses) For every a F, there exists some b F such that
a + b = 0.
(Multiplicative Inverses) For every a F , except a = 0, there exists some
c F such that a c = 1.

4.2

Polynomial

Mathematical definition of polynomial:

Suppose there are n variables, x1 , x2 , . . . , xn . A monomial over a field F is a
product of the form
xd11 xd22 xdnn ,

in which all of the exponents d1 , d2 , . . ., dn are positive integers. The total

degree of the monomial is d1 + + dn . A polynomial f in variables x1 , . . . , xn
with coefficients in F is a finite linear combination of monomials that have
coefficients in F. A polynomial can be expressed as
m
X
ci mi ,
i=1

in which mi is a monomial as shown in (4.47), and ci F is a coefficient. If

ci 6= 0, then each ci mi is called a term.
Polynomial F (x1 , x2 , ..., xn ) is actually a group with respect to addition operation.
Affine space of dimension n is defined as follow:
Fn = {(c1 , . . . , cn ) | c1 , . . . , cn F}.
A polynomial f F[x1 , . . . , xn ] can be converted to a function that maps F n
to F : f : Fn F,

4.3

Varieties

Let F be a field and let {f1 , . . . , fk } be a set of polynomials in F[x1 , . . . , xn ].

The set V (f1 , . . . , fk ) = {(a1 , . . . , an ) F | fi (a1 , . . . , an ) = 0 for all 1 i k
is called the (affine) variety defined by f1 , . . . , fk . One interesting fact is that
unions and intersections of varieties are varieties.
Intersection of two varieties f and g are given as the affine variety defined by
both f,g. V (f1 , . . . , fk ) V (g1 , . . . , gl ) = V (f1 , . . . , fk , g1 , . . . , gl )
To obtain union of varieties, we just need to multiply the polynomial.