Its hard to install dumb waiters in most hospitals because of the expense and time of construction. The nice thing about the robot is that it can move horizontally (through the hallways) and vertically (using the elevator) and be installed rather quickly.
The robot does not evolve or learn from its environment. So, in a sense, it does not use AI. Just good old fashioned servoing algorthims through a goal point established with its sensor fabric. Still, its pretty cool to watch it move.
The robot is not built or sold to replace nurses in any way. Its sold on the proposition that it helps the skilled care givers spend more time on thier primary tasks - thier patients.
The baritone voice was from a professional voice model. It does not sound like Darth Vader. It kind of sounds like a game show host. We also make robots that have a female voice. The gender of the voice is chosen by the site. Future versions may have custumizable voices.
Re:what about the human side
on
Hospital Robots
·
· Score: 1
The big focus of the robot's marketing package is that it helps nurses stay with the patients instead of running to central supply to pick up this or that. It has been really useful for night shifts where it is hard to find good people (honest, willing to work through the night, etc). The focus has NOT been to get between the nurse/doctor and the patient.
The article was short on technical details like a lot of broad interest articles are. But the current version of the robot uses a good ole' 68K as its main processor and various 68hc11's as slave processors. It has several systems that work together to bring about its behavior. Navigation and path planning are done in the 68K while the lower level stuff like sonar ranging and collision detection are done with the 68HC11's. There is a 486 (or 286, depending on the version) in the current incarnation that facilitates Ethernet connectivity and another for structured vision to detect obstacles in front of the robot.
Given the relative simplicity of the robot's architecture it navigates really well and one of its biggest problems is crowded hallways. People are moving constantly and the robot cannot currently infer where the detected obstacle has moved. So most people do not have it plan its route through crowed hallways.
I work for Pyxis, but the usual disclaimers apply - the views expressed here are my own and I am trying not to share too much info because a lot of it is proprietary.
-George
Slashdot Mirror
Its hard to install dumb waiters in most hospitals because of the expense and time of construction. The nice thing about the robot is that it can move horizontally (through the hallways) and vertically (using the elevator) and be installed rather quickly.
The robot does not evolve or learn from its environment. So, in a sense, it does not use AI. Just good old fashioned servoing algorthims through a goal point established with its sensor fabric. Still, its pretty cool to watch it move.
Many of the senior volenteers that I have met like the robot.
The robot is not built or sold to replace nurses in any way. Its sold on the proposition that it helps the skilled care givers spend more time on thier primary tasks - thier patients.
The baritone voice was from a professional voice model. It does not sound like Darth Vader. It kind of sounds like a game show host. We also make robots that have a female voice. The gender of the voice is chosen by the site. Future versions may have custumizable voices.
The big focus of the robot's marketing package is that it helps nurses stay with the patients instead of running to central supply to pick up this or that. It has been really useful for night shifts where it is hard to find good people (honest, willing to work through the night, etc). The focus has NOT been to get between the nurse/doctor and the patient.
The article was short on technical details like a lot of broad interest articles are. But the current version of the robot uses a good ole' 68K as its main processor and various 68hc11's as slave processors. It has several systems that work together to bring about its behavior. Navigation and path planning are done in the 68K while the lower level stuff like sonar ranging and collision detection are done with the 68HC11's. There is a 486 (or 286, depending on the version) in the current incarnation that facilitates Ethernet connectivity and another for structured vision to detect obstacles in front of the robot. Given the relative simplicity of the robot's architecture it navigates really well and one of its biggest problems is crowded hallways. People are moving constantly and the robot cannot currently infer where the detected obstacle has moved. So most people do not have it plan its route through crowed hallways. I work for Pyxis, but the usual disclaimers apply - the views expressed here are my own and I am trying not to share too much info because a lot of it is proprietary. -George