**5.3 Speech recognition**

Speech recognition is also based on the monitoring of human motions. When the wearable electromechanical is attached on the throat, it could record muscular movements in order to collect and recognize speech sounds. This is permitted by the fact that the throat muscle exhibits different degrees of stretching or shrinking strains when speaking different words. Due to the tiny changes caused by throat motion, the strain sensor used in speech recognition should have high sensitivity. The GF of GWF strain sensor can be as high as 103 with 2–6% strains, 106 with higher strains (>7%), and ~35 with a minimal strain of 0.2%, which is suitable for this application. The results show test signal waveforms of all 26 english letters [70]. As expected, the waveforms are unique and repeatable for all letters. Since each

individual speech organ is different, people can easily distinguish whether a given voice comes from the same person. This demonstrates that wearable electromechanical can be used in speech recognition.

## **5.4 Human-machine interface**

Human-machine interfaces and robotic remote controlling are greatly beneficial in surgery or a highly risky work that requires the replacement of robotics. The electromechanical sensor used in human-machine interface is typically mounted on body joint, which are normally bended or stretched at large degree of deformations; thus, high stretchability (>50%) is required. The robotic controlling is demonstrated in **Figure 7**, and the wearable strain sensors are based on the hybrid of polyaniline and gold nanowires for a smart glove [9]. The sensor-based-smart glove is used to control the movement of a robot through wireless signals (**Figure 7a**). The robot is at relaxed state (a1) and works as an arm that can clamp (a2), lift up (a3), put down (a4), and release (a5) an object based on different postures of human fingers as wearing the sensor. **Figure 7b** reveals the remote control on the robot movement by a strain sensor based on graphene. As can be seen in this figure, b1 and b4 demonstrate the robot at the relaxed state. As the strain sensor is stretched or bended, the robot starts working (b2 and b5) and moves to the controller (b3 and b6).
