Robots are hardware controlled by software. Some of the recent technologies enabling robotics to advance are:

Robotic Technologies of Interest

AI Awareness
Awareness of surroundings by localization and mapping of the environment, a semantic understanding of objects in the environment, and the ability to choose appropriate actions.
AI Methods
AI Methods
-Simultaneous Localization and Mapping (SLAM) is now a proven technology in limited environments. -Semantics are limited by the training resources required. -Planning and Decision algorithms are working in ever more complex environments.
Awareness
Deep Learning
Deep learning in neural networks are taking on increasingly complex situations, e.g. finding an effective path in real time with other moving object in the environment.
AI Capability
AI Capability
AI will be limited by computing resources and sensor information for many years to come. There are many tasks within the current AI limit that are valuable.
Deep Learning
Human Interaction
It has been documented that acceptance of robots' is strongly correlated to the robot's sophistication in subtle communications cues. The human-robot interaction models are starting to show results enhancing the robots sophistication.
Perception of Intent
Perception of Intent
Humans expect to be able to communicate subtly. Robot perception is being developed to understand even the expression of emotion as a indication of human intent (e.g. raising your eyebrows).
Human Interaction
Robot Hands
There are many times more different end effectors (a robot’s hand) than there are robots, because holding different objects require unique grips. Examples of end effectors include: vise jaws, flexible fluid filled 'fingers', suction cups, sensor controlled 'finger', and many more.
Manipulation Challanges
Manipulation Challanges
End effectors that hold arbitrary objects have limitations with strength and abrasion. Common jaws end effectors will damage even soft metal objects unless specifically designed and controlled.
Robot Hands