TY - GEN
T1 - Intelligent scanning collision avoidance device with risk assessment
AU - Haddad, Malik
AU - Shamieh, Jamal
AU - Sanders, David
AU - Gharavi, Amir
AU - Tewkesbury, Giles
AU - Hassan-Sayed, Mohamed
AU - Langner, Martin
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
PY - 2024/8/1
Y1 - 2024/8/1
N2 - This paper presents an Intelligent Scanning Collision Avoidance Device (Intelligent-SCAD) which is used to detect obstacles in a powered wheelchair surroundings to avoid related risk consequences. The Intelligent-SCAD provides a safe direction for the wheelchair. Inputs to the Intelligent-SCAD originate from a single rotating ultrasonic transducer fixed to the wheelchair. Readings from the ultrasonic transducer are used to train and test different Artificial Intelligence (AI) algorithms. The AI algorithms used were: Artificial Neural Network, Decision Tree, optimised Tree and optimised K-Nearest Neighbour. An algorithm is selected based on a compromise between accuracy and complexity. The optimised K-Nearest Neighbour algorithm provided the highest testing accuracy and relatively straightforward operation when compared with the other algorithms used. The new device applies optimised K-Nearest Neighbour to predict a safe direction for a wheelchair. The user can override the new system if necessary.
AB - This paper presents an Intelligent Scanning Collision Avoidance Device (Intelligent-SCAD) which is used to detect obstacles in a powered wheelchair surroundings to avoid related risk consequences. The Intelligent-SCAD provides a safe direction for the wheelchair. Inputs to the Intelligent-SCAD originate from a single rotating ultrasonic transducer fixed to the wheelchair. Readings from the ultrasonic transducer are used to train and test different Artificial Intelligence (AI) algorithms. The AI algorithms used were: Artificial Neural Network, Decision Tree, optimised Tree and optimised K-Nearest Neighbour. An algorithm is selected based on a compromise between accuracy and complexity. The optimised K-Nearest Neighbour algorithm provided the highest testing accuracy and relatively straightforward operation when compared with the other algorithms used. The new device applies optimised K-Nearest Neighbour to predict a safe direction for a wheelchair. The user can override the new system if necessary.
KW - Artificial intelligence
KW - Collision avoidance
KW - Disabled
KW - Risk
KW - Steer
KW - Wheelchair
UR - https://www.scopus.com/pages/publications/85201121576
U2 - 10.1007/978-3-031-66336-9_8
DO - 10.1007/978-3-031-66336-9_8
M3 - Conference contribution
AN - SCOPUS:85201121576
SN - 9783031663352
T3 - Lecture Notes in Networks and Systems
SP - 112
EP - 123
BT - Intelligent Systems and Applications - Proceedings of the 2024 Intelligent Systems Conference IntelliSys Volume 4
A2 - Arai, Kohei
PB - Springer
T2 - Intelligent Systems Conference, IntelliSys 2024
Y2 - 5 September 2024 through 6 September 2024
ER -
