Page 252 - 2024-Vol20-Issue2

248 |
and specific challenges of healthcare settings. GA-PRM algo-
rithm is attempting to fill the gap of providing a robust and
adaptable path-planning solution that can navigate healthcare
environments efficiently and safely. Dynamic obstacles, the
need for smooth and precise movements, and the requirement
for patient safety characterize healthcare environments.
III. METHODOLOGY
In this section, the methodology employed for the develop-
ment and implementation of the novel path-planning algo-
rithm is discussed. A comprehensive description of the algo-
rithm is provided, clarifying its complexities and fundamental
components. Core techniques utilized in the approach are
detailed, underscoring their essential role in effective path-
planning. The seamless integration of genetic algorithms
and probabilistic roadmaps is explained, displaying how this
combination enhances the algorithm’s performance and effi-
ciency. Modifications or enhancements to existing algorithms
are highlighted, spotlighting how these innovations and ad-
vancements refine the approach.
A. Techniques
The GA-PRM algorithm is a path-planning algorithm that
combines the strengths of two other algorithms: Genetic Al-
gorithm (GA) and Probabilistic Roadmaps (PRM).
The GA component of GA-PRM is inspired by natural
selection. It uses a population-based search approach to iter-
atively evolve a set of candidate paths. Each candidate path
is represented as a string of waypoints in the robot’s configu-
ration space. The fitness of each candidate path is evaluated
based on its length and how well it avoids obstacles. The fittest
individuals from each generation are selected to contribute to
the next generation, which helps the algorithm converge to
optimal solutions.
The PRM component of GA-PRM constructs a roadmap
of the workspace. The roadmap is a graph that represents
all the feasible paths between random configurations in the
workspace. Random configurations are sampled uniformly
across the workspace, and collision checking ensures that
they avoid obstacles. The roadmap is built by connecting
configurations through collision-free edges.
The GA-PRM algorithm leverages the strengths of both
the GA and PRM algorithms. The GA component enables the
algorithm to search for optimal paths in a large and complex
search space, while the PRM component ensures that the
algorithm finds feasible paths even in cluttered environments.
The combination of these two algorithms results in a path-
planning algorithm that is both efficient and effective.
The core components and techniques used in the algorithm
are as follows:
Sabeeh & Al-Furati
1. Roadmap generation: A roadmap is a graph that repre-
sents the possible paths that a robot can take in an envi-
ronment. Randomly sampling configurations within the
workspace and then connecting them with collision-free
paths generate it. The roadmap construction accounts
for the presence of static obstacles, ensuring that the
robot can traverse through open paths.
2. Genetic algorithm: The genetic algorithm is a potent
tool for optimizing robot path planning. It iteratively
improves the path by evolving potential routes through
generations. It starts with a population of path segments,
each representing a possible way for the robot to reach
its goal while avoiding obstacles. These segments are
evaluated based on how well they work. Parents are
selected from the population based on their effective-
ness and used to create new paths through crossover
and mutation. Crossover involves parents exchanging
parts of their paths to make new ones, while mutation
makes small changes to paths. This process continues
over generations until a suitable path is found, allowing
the robot to reach its goal efficiently.
3. Handling dynamic obstacles: To ensure robots nav-
igate safely in dynamic environments, there are two
key methods: potential field-based path-planning and
genetic algorithms. Potential field-based planning cal-
culates forces guiding the robot towards its goal and
away from obstacles, allowing it to adapt its path in
real-time as conditions change. Genetic algorithms, on
the other hand, use iterations to find the best paths by
considering various factors like distance and efficiency,
making them suitable for navigating complex spaces
with both static and moving obstacles. For instance, in
a busy hospital, robots can first use potential field-based
planning to avoid patients and staff, and then optimize
their route with genetic algorithms for safe and efficient
navigation. In healthcare settings, ultrasonic sensors
can complement infrared (IR) sensors by detecting vari-
ous types of moving objects and equipment, ensuring
collision-free movement.
4. Path smoothing: Path smoothing is a crucial part of
path-planning, especially in complex and dynamic set-
tings. It reduces noise and uncertainties caused by sen-
sors and moving obstacles, making the path smoother
and more reliable. Kalman filtering is a statistical tech-
nique used for this purpose. It estimates the robot’s
position and speed by combining previous estimates
and noisy measurements. This process repeats to refine
the estimate, resulting in a more accurate path.
5. Performance metrics: The performance of the new