Path Planning & Navigation

Dr Akhil V M
Asst. Professor,
Department of Mechanical Engineering,
Amrita Vishwa Vidyapeetham,
Amritapuri Campus 1
What is a path ?

2
Any important in path
planning ?

3
Path planning and trajectory planning

• Path planning and trajectory planning are crucial issues in the field of Robotics and, more
generally, in the field of Automation.
• Indeed, the trend for robots and automatic machines is to operate at increasingly high
speed, in order to achieve shorter production times.
• The high operating speed may hinder the accuracy and repeatability of the robot
motion, since extreme performances are required from the actuators and the control
system.
• Care should be put in generating a trajectory that could be executed at high speed, but at
the same time harmless for the robot, in terms of avoiding excessive accelerations of the
actuators and vibrations of the mechanical structure.

4
5
6
 Path planning is the process of finding an optimal
or feasible path between a start and a goal
position, while avoiding obstacles and respecting
constraints.

Path planning is a merely geometric matter,

Path planning

because it is defined as the generation of a

geometric path, with no mention of any specified
time law.

On the other hand, trajectory planning consists in

assigning a time law to the geometric path

7
8
9
10
11
12
13
14
15
Shortest distance between 2 point

16
17
 • Path planning is a crucial component in many applications
such as robotics, autonomous vehicles, drones, video games,
and virtual reality.
• Robotics: Path planning is used extensively in robotics to plan
the movement of robots in a given environment. This includes
industrial robots in manufacturing, mobile robots in
warehouses and hospitals, and autonomous robots in outdoor
Applications environments.
• Autonomous Vehicles: Path planning is an essential
of Path component of autonomous vehicles, such as self-driving cars,
trucks, and drones. Path planning algorithms are used to plan
Planning the route of the vehicle while avoiding obstacles and ensuring
safe driving.
• Video Games: Path planning is used in video games to enable
non-player characters (NPCs) to navigate the game world. This
includes games such as first-person shooters, role-playing
games, and strategy games.

18
• Virtual Reality: Path planning is used in virtual reality to enable
users to move around in a virtual environment. This includes
applications such as virtual training simulations, architectural
visualizations, and gaming.
• Motion Planning: Path planning is used in motion planning for
mechanical systems such as robotic arms, CNC machines, and
automated assembly lines.
• Path Optimization: Path planning is used in optimization
problems such as finding the shortest path or the fastest path
for a vehicle or agent to travel.
• Agriculture: Path planning is used in precision agriculture to
plan the movement of autonomous tractors and drones for
efficient crop management.
• Disaster Response: Path planning is used in disaster response
applications to plan the movement of robots and drones for
search and rescue operations in hazardous environments.

19
Types of Path Planning
Algorithms

• There are several types of path

planning algorithms, including
heuristic search, probabilistic
sampling, potential fields, and
machine learning-based methods.

20
Heuristic Search
Algorithms
• Heuristic search algorithms are a family of
algorithms that use heuristic functions to guide
the search process towards the goal.
• The heuristic function is an estimate of the cost
from the current node to the goal node.
• The algorithm tries to expand the nodes with
the lowest estimated cost first to reach the goal
state efficiently.
• Here are some common heuristic search
algorithms used for path planning: Examples
include A* search, Dijkstra's algorithm, and
breadth-first search
21
• A* Search: A* search is a widely used algorithm
that combines the benefits of uniform-cost
search and greedy search.
• It expands the node that has the lowest sum of
the cost of the path so far and the estimated cost
to the goal node.
• A* is guaranteed to find the optimal path if the
heuristic function is admissible, which means
that it never overestimates the true cost to the
goal

22
23
24
• Dijkstra's Algorithm: Dijkstra's algorithm is similar to
A* search, but it does not use a heuristic function.
Instead, it expands the node with the lowest cost to
reach it from the start node. Dijkstra's algorithm is
guaranteed to find the optimal path if all edge costs
are non-negative.

25
26
• Greedy Best-First Search: Greedy
best-first search is a heuristic search
algorithm that expands the node that
has the lowest estimated cost to the
goal node. It does not consider the
cost of the path so far, which makes it
faster but may not find the optimal
path.

27
• Hill Climbing: Hill climbing is a local search algorithm that starts at a random node and
moves to the neighboring node with the lowest estimated cost. It is fast but may get
stuck in a local optimum and not reach the global optimum.
• Beam Search: Beam search is a variant of best-first search that expands a fixed number
of nodes with the lowest estimated cost. It can be faster than best-first search but may
miss the optimal path if the beam width is too narrow.
• Iterative Deepening A*: Iterative deepening A* is a variant of A* search that performs
multiple depth-limited searches with increasing depth until the goal is found. It has the
same optimality guarantee as A* search but uses less memory.

28
29
Probabilistic
Sampling
Algorithms
• Probabilistic sampling algorithms
randomly sample the
configuration space to find
feasible paths. Examples include
Rapidly-exploring Random Trees
(RRT) and Probabilistic Roadmaps
(PRM).

30
Potential Fields
Algorithms

• Potential fields algorithms model the environment

as a potential field and use gradients to guide the
path. Examples include artificial potential fields
and harmonic functions.

31
Potential fields algorithms are a family of path planning algorithms that use the
concept of potential fields to plan a path for a robot or agent in a given environment.

In potential fields algorithms, the environment is modeled as a field of forces, where

the robot is attracted to the goal location and repelled from obstacles.

The resulting path is a trajectory that minimizes the total energy required to move
from the starting location to the goal location.

Here are the main types of potential fields algorithms:

32
Artificial Potential Fields (APF): APF is a
widely used potential fields algorithm
that represents the environment as a
field of attractive and repulsive
potentials.
• The robot is attracted to the goal
location and repelled from obstacles,
which generates a resultant force that
directs the robot towards the goal.
• APF is simple to implement and
computationally efficient, but it can
suffer from local minima and
oscillations around obstacles.

33
• Dynamic Window Approach (DWA): DWA is a
variant of potential fields that uses a dynamic
window of allowable velocities and computes a
safe and feasible trajectory for the robot.
• DWA considers the robot's kinematics and
dynamics constraints to generate a smooth and
collision-free path.
• It is well suited for mobile robots that operate in
dynamic environments.

34
 Overview of DWA.

M. Kobayashi and N. Motoi, "Local Path Planning: Dynamic Window Approach With Virtual Manipulators
Considering Dynamic Obstacles," in IEEE Access, vol. 10, pp. 17018-17029, 2022, doi:
10.1109/ACCESS.2022.3150036.
35
• Elastic Band (EB): EB is a path
planning algorithm that uses a
band of virtual elastic links to
guide the robot towards the goal.
• The elastic band is a one-
dimensional curve that is attracted
to the goal location and repelled
from obstacles.
• The robot's position is optimized
by deforming the elastic band
through an iterative process until
the goal is reached.

36
 Framework of the EB and the EB-RRT algorithm. (a) EB framework. (b) rapidly exploring
random tree (EB-RRT) framework.
J. Wang, M. Q. . -H. Meng and O. Khatib, "EB-RRT: Optimal Motion Planning for Mobile Robots," in IEEE Transactions on
37
Automation Science and Engineering, vol. 17, no. 4, pp. 2063-2073, Oct. 2020, doi: 10.1109/TASE.2020.2987397
• Vector Field Histogram (VFH):
VFH is a potential fields algorithm
that represents the environment
as a polar histogram of obstacle
density around the robot.
• The robot selects a heading
direction that maximizes the free
space in the histogram and avoids
collisions with obstacles.
• VFH is computationally efficient
and can handle complex
environments with multiple
obstacles.

38
 Machine
Learning-based
Algorithms
• Machine learning-based algorithms
use neural networks or other learning
techniques to learn the optimal path.
Examples include Deep Reinforcement
Learning and Convolutional Neural
Networks.

39
Challenges in Path Planning
• Path planning can be challenging due to issues such as high-
dimensionality, non-convexity, uncertainty, and dynamic
obstacles.
• Path planning is a complex problem that involves several
challenges. Here are some of the main challenges in path
planning:
• High Dimensionality: Path planning problems in high-
dimensional spaces, such as robotic arms, can have an
exponentially large search space, making it challenging to find an
optimal path.
• Dynamic Environments: Path planning in dynamic environments
with moving obstacles and unpredictable events is challenging.
The robot must continuously adapt its path to avoid collisions
and reach the goal safely.
• Uncertainty: Path planning in uncertain environments with
incomplete or noisy sensor data can lead to suboptimal paths or
collisions. Robust path planning methods that can handle 40
uncertainty and adapt to changing conditions are needed.
• Real-Time Constraints: In many applications, such as autonomous vehicles or
drones, path planning must be done in real-time, which imposes time constraints
on the search algorithm. Fast and efficient path planning methods that can operate
in real-time are necessary.
• Local Minima: Path planning algorithms can get stuck in local minima or oscillate
around obstacles, resulting in suboptimal or inefficient paths. Global optimization
methods that can escape local minima and find the global optimum are needed.
• Human-Robot Interaction: In some applications, such as collaborative robotics or
social robots, path planning must take into account human preferences, safety, and
comfort. Path planning algorithms that can generate socially acceptable paths and
interact with humans are necessary.

41
42
Trajectory planning in robotics

• Trajectory planning is the process of determining a path for a robot to follow

in order to complete a task or reach a desired goal.
• It involves taking into account the robot's physical capabilities and limitations,
as well as any obstacles or constraints in the environment.
• Trajectory planning can be done using a variety of techniques, including
kinematic and dynamic models, optimization algorithms, and machine
learning.
• The goal of trajectory planning is to minimize the time and energy required for
the robot to complete its task, while also ensuring safety and accuracy.

43
44
45
46
47
48
49
Trajectory planning is a crucial aspect of robotic motion planning,
and it involves generating a path that a robot can follow from its
current location to a desired goal location.

• Kinematic constraints: These are constraints that are related to

the robot's physical limitations. For example, the robot's arm
may have a limited range of motion or a maximum velocity it can
achieve.
• Dynamic constraints: These constraints are related to the
robot's dynamics, such as the maximum acceleration or
deceleration that the robot can undergo. These constraints
ensure that the robot moves in a smooth and stable manner.
• Environmental constraints: These constraints are related to the
environment in which the robot operates. For example, the
robot may need to avoid obstacles or stay within a defined
workspace.

50
• Collision avoidance: This constraint is related to
ensuring that the robot does not collide with any
objects in its environment. Collision avoidance is
critical for ensuring the safety of the robot and the
surrounding environment.
• Operational constraints: These are constraints
related to the operational requirements of the task
being performed by the robot. For example, the
robot may need to move quickly or slowly, or it may
need to follow a specific trajectory to complete the
task.
• Time constraints: These constraints are related to
the time required for the robot to complete a task.
For example, the robot may need to complete a
task within a specific time frame.

51
Almost every technique found in the scientific literature on the
trajectory planning problem is based on the optimization of some
parameter or some objective function.
The most significant optimality criteria are:
(1) minimum execution time
(2) minimum energy (or actuator effort)
(3) minimum jerk

Gasparetto, Alessandro, and V. Zanotto. "A new method for smooth trajectory planning of robot
manipulators." Mechanism and machine theory 42, no. 4 (2007): 455-471.
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
Kamali, Kaveh, Ali Akbar Akbari, and Alireza Akbarzadeh. "Trajectory generation and control of
a knee exoskeleton based on dynamic movement primitives for sit-to-stand
74
assistance." Advanced Robotics 30, no. 13 (2016): 846-860.
75
References

• LaValle, S. M. (2006). Planning algorithms. Cambridge university press.

• Thrun, S., Burgard, W., & Fox, D. (2005). Probabilistic robotics. MIT press.
• Karaman, S., & Frazzoli, E. (2011). Sampling-based algorithms for optimal motion planning. The international journal of robotics research, 30(7), 846-894.
• Choset, H. (2005). Principles of robot motion: theory, algorithms, and implementations. MIT press.
• Borenstein, J., Everett, H. R., & Feng, L. (1996). Where am I? Sensors and methods for mobile robot positioning. Robotics and Autonomous Systems, 17(3), 231-245.
• Minguez, J., Montano, L., Grau, A., & Reina, A. J. (2004). Reactive navigation based on a pure reactive potential field and path planning. Robotics and Autonomous
Systems, 47(1), 55-67.
• Khatib, O. (1986). Real-time obstacle avoidance for manipulators and mobile robots. The International Journal of Robotics Research, 5(1), 90-98.
• Fox, D., Burgard, W., & Thrun, S. (1997). The dynamic window approach to collision avoidance. IEEE Robotics & Automation Magazine, 4(1), 23-33.
• Bekey, G. A. (2005). Autonomous robots: from biological inspiration to implementation and control. MIT press.
• Fraichard, T., & Asama, H. (2004). Motion planning for mobile robots: a survey. IEEE Transactions on Robotics and Automation, 20(2), 195-202.

76
77