Table 5.

Comparison of human and robot interaction with relevant research methods.

Reference Papers	Sensory Approach		Algorithm	Hardware	Number of Postures/Position	Pros	Accuracy	Cons
	Method	Fusion
Q. Hu et al. 2021 [20]	1024 sensors Pressure sensor	Yes	HOG, SVM, and CNN	Arduino Nano and CPU	6	<400 ms, sampling and processing	86.94% to 91.24%	Contact approach
Matar et al. 2020 [37]	1728 FSR sensors	Yes	HOG + LBP, FFANN	CPU	4	Health monitoring	97%	More usage of sensors
R. Tapwal et al. 2023 [38]	Two flex force sensors	Yes	K-means	Arduino Uno and CPU	4	Health monitoring	~99.3%	consumes 17.5 W, contact approach
Hu, D et al. 2024 [39]	32 Piezoelectric sensor	Yes	S3CNN	N/A	4	Effectively detects nuanced pressure disturbances	93.0%	not applicable
Y. Tanaka et al. 2020 [40]	Camera	No	Amygdala	FPGA	_	Interaction with subject based face recognition	>90%	not applicable
T.Kim et al. 2024 [41]	Multi sensors and Camera	Yes	DFS, 3D routes, Vision algorithms	Intel NUC and Nvidia Jetson Xavier		Multi floor service and mapping	N/A	Costlier in implementation
Proposed	10 Ultrasonic sensors	Yes	Human Localization, Triangulation Navigation algorithm	FPGA		Parallel computing, <200 ns, sampling, and computation. Adaptive localization-based robot services	98.4%	PR flow will be preferred in future usage