Robotics systems have garnered a lot of respect and attention in the age of artificial intelligence. Systems that replicate human behaviour thought processes, and other everyday problem-solving abilities are known as humanoids. The creation of bipeds is a difficult endeavour that involves both hardware costs and technological considerations. This article focuses on recent developments in the field of robotics, particularly humanoids, their design as systems, ethical issues, how they interact with Society, applications, and upcoming difficulties related to them. We have developed (i) a new extensive design cycle for robots (EDLC) framework to develop humanoid and (ii) machine learning framework to classify activity recognition. Design cycle will help to design robot in modular fashion while activity recognition will further help to recognize Robot-Human(R-H Model),Robot-Robot(R-R Model) interaction in gaming, social environments, surveillances and Biometric application developments. Results shows that Machine learning (ML) is useful in classifying human activities. SVM (Support Vector Machine) Has shown accuracy with 95.65% on UCI activity dataset while Linear Regression (LR) has shown better accuracy with 97.74%.