EMIEW2 was developed as part of Hitachi's efforts to create a service robot with diverse communication functions that could safely coexist with humans while conducting necessary services.o ensure agility and safety in an office environment, EMIEW2 was designed to a compact height of 80 cm and a portable weight of 14 kg. Further, to enable harmonious interaction while working with people, EMIEW2 is able to travel at the same speed as humans as well as stop stably. The 2-wheel 6 km/hr high speed feature developed for EMIEW1 is retained in a two 2-wheel mechanism, which enables agile 2-wheel movement, and a wheeled-leg type mechanism was developed to get over floor level differences in the office.

A service robot that work in a office building is required to run over safely an obstacle on a floor, such as a threshold, mattress, or electric cables, therefore, "Active suspension" and "Idling control technology" were built into the leg mechanism of EMIEW 2. "Active suspension" is consisted of a spring and an actuator. The spring absorbs softly a big impact generated by running over a step on floor, and the actuator recovers the inclination of the robot according to the transformation of the spring. Moreover, "Idling control" is a function to control an excessive idling of the wheel by the uncontrollability in the wheel jumping timing when the robot run over the step, and for a steady running to continue after it lands.
To talk with the person in a real office building, the robot is required to recognize accurately human voices with serious noisy condition in which the noises come from upper and lower, right and left, various directions, moreover the robot mechanism becomes a source of the noise. This time, the technologies that suppress the noise efficiently with 14 microphones installed in the head of EMIEW 2 were developed. One of them is a horizontal and vertical voice direction analyzing technology that is able to focus the voice search space efficiently, and another is noise-canceling technology against the noises generated by the robot self. The voice recognition performance has been improved rapidly by these technologies.