| dc.description.abstract | Robots significantly improve productivity and production efficiency due to process
optimisation and low production costs. Job enrichment and fulfilment can be achieved
through improved workflows and role distribution. This will also lead to an improved
capacity to handle complex assignments, perform tedious and sophisticated tasks
quickly, enhance workers' safety, and improve the customization of goods and
services. The adoption of robots in Kenyan small and medium-sized enterprises
(SMEs) manufacturing industries has been gradual and faces many challenges. Key
among the challenges is the few skilled labourers with robot programming capabilities
for the varying manufacturing environments. This has led to the need for more
competitiveness in the manufacturing sector with other countries, especially on the
global stage. This can be immensely magnified in flexible manufacturing systems,
especially when switching product types to robotised systems requires higher costs and
time. The inferior skill set of people interacting with the robots warrants designing and
generating
user-friendly
programming
approaches
using
kinesthetic
teaching and augmented and virtual reality. The Kenyan government's development
agenda aims to achieve the 2030 goals using emerging innovative technologies,
including robotics, machine learning, and artificial intelligence (AI). This research
purposed to analyse the demonstrative-kinesthetic teaching (DKT) approach to robotic
manipulators for efficient material handling applications. The robotic arm was
programmed using structured texts and DKT to determine coordinate configurations,
the desired position's accuracy, and the DKT's efficiency. A control platform was
created using Visual Studio to allow the arm to be programmed demonstratively using
the lock arm button. This allowed the arm to record the demonstrations while the user
did the programming. Palletizing and contour path welding experiments were
conducted to validate the study and collect the requisite data. Structured texts were
used as the control for the experiment. The results found that the mean inverse
kinematics were the same for both methods at the α=0.05 significance level, F=0.03,
P=0.86, and Fcritical =5.98. Joint 2 had a low percentage error at 2.12 % and a high for
joint 4 at 5.24%, majorly due to user level of accuracy. The DKT had an 80% and
66.67% efficiency on experimental time for palletising and contour path welding,
respectively, compared to structured text. The conclusion was that DKT provided a
means of finding the joint configurations in concurrence with analytical solutions. The
robotic manipulator was able to trace paths and desired positions accurately. DKT
provided more accessible programming for non-skilled floor operators than structured
texts. Some recommendations were the inclusion of wearable devices in the DKT
approach and shifting the control platform created to universal set-up platforms. | en_US |
