Tuesday, September 16, 2014
The SR4 Office Robot
SR4 Office Robot
Walk through this near-future scenerio

I have a project folder I wish to send to Kate Rahn, who works up on the next floor on the other side of the building. I hit the "Robot Pool" Icon on my desktop computer and up pops a task form, pre-filled with my own desk ID and name. I key in the end-destination I will want, check a task selection (in this case, "Delivery") and press "Submit". If I donít happen to know Kateís desk ID by heart, I can key her name (or enough for the PC to recognize who I am talking about - or I can browse the employee directory for her destination ID).

In a moment, up pops a message that indicates that robot R17 is on its way to my desk and gives me an estimated time of arrival (ETA). "Great", I exclaim silently to myself - R17 is one of my favorites; painted Ďround its body with a scene from the Rocky Mountains.

Shortly, R17 wheels up to my desk and announces itself in a soft robotic voice. I place the project folder in the robot's carry-tray, select the "GO" button on the robot task form on my PC and off the robot scoots to make it's way to Kate's desk. In the meantime, a message appears on Kate's PC that R17 is on its way, sent by me, and with an ETA countdown-minutes digital-indicator beside the message.

Following internally stored facility-maps of the building, the robot tracks the shortest path to the elevator, calls and operates the elevator digitally, avoids people and obstacles (objects not on the maps), and shortly arrives at Kate's desk. Kate's not there. The robot waits patiently for an appropriate period of time, dispenses its carry-tray's contents into Kate's in-basket and then heads for the nearest robot pool location.

Robots also grab coffee and snacks from the Starbuck's cafe on the first floor, deliver and/or retrieve copies to and from the copy center, deliver mail from and to the mail-room, and generally handle those chores that otherwise are just time-burners for me and others.

Random Running

Sometime in early 2012, a scaled up version of the SR4-N (Navigator) robot (between four and four Ĺ feet tall and with industrial grade wireless network range and performance) will begin executing a series of tests we call "random running" in a major office building in Manhattan. The robot will be loaded with a facilities map of a single floor of the building which will feature a series of waypoints (navigation node points) noted on the map. The waypoints collectively define the open spaces that the robot is able to reach or pass through without getting stuck in spaces, corridors or pathways that are two small for the robot.

Random Running entails the robot, which is coordinate aware and which can read and interpret the map, to (1) select a destination waypoint at random anywhere in the current or adjacent room or hallway, (2) plot the most direct path (a series of straight lines and specified turns), to the destination point and then (3) follow the pathway until the robot arrives at the destination point. Once the robot arrives at the point chosen, it repeats the process again, choosing a new destination point at random, etc. This process is repeated continuously for a period of time and/or for a finite number of trips as a burn-in exercise of (1) the robot, (2) the map, (3) the officeís "hot spot" network, and (4) the command and control console that continuously tracks the robotís whereabouts, performance and other vital signs.

The robot tracks its progress carefully by watching its wheel encoders which provide the feedback it needs to track and control its progress along each path segment and through each turn. Periodically the robot takes a fix on beacons mounted in the space it is moving through, in order to triangulate its position and correlate that computation with where the robot already believes itself to be. When necessary, the robot adjusts its bearing and time-of-travel to compensate for any distance it may have been found via the beacons to be off from its computed path.

The Random Running task is designed to have the SR4-N robot run for a number of hours without any control center intervention and testing the robotís ability to handle mission after mission autonomously and without any tracking error beyond the acceptable limits of two to three inches. This error limit applies both to the robotís adherence to path throughout its entire trip, as well as to the accuracy of its achievement of target positions.

The SR4-N Office Robot Prototype Application

Smart Robots' Office Robot prototype application is built upon the foundation of the SR4 Robot's map-tracking navigation system and the robot's on-board web-server and wireless networking technology. The robot operates in an office building equipped throughout as a wireless "hot spot" and with its own experimental local internal GPS system.

It is anticipated that the Random Running test described below is going to be achieved by one of two methods over the next 12+ months; either (1) with an SR4-N configuration of the SR4 which relies on whatever version of Smart Robotsí experimental Triangulator system is available at that time, or (2) with an SR4-B or P or Pb configuration of the SR4 which relies on a third party outside referencing system for the augmentation of its self-tracking. The SICK laser tracker is one such third party system.

Accurate navigation for the office application is principally based on (1) a motor-wheel assembly and control system that can deliver routine accurate dead-reckoning (the proper term is actually "ded" reckoning, for "deduced") and tracking for the robot, and (2) an outside referencing system to regularly provide course-validation-and-correction information to the robot, in case its wheels are thrown off track or the robot is otherwise propelled off course.

Join the Effort

Join the effort of developing this first generation of distributed point-to-point office robots, either as an educational project or in preparation for you or your organization or firm to become a player in this exciting new application of mobile robots over the next several years. An earlier generation of mobile office robots has existed for over 30 years in the form of machines that follow chemical and/or embedded metal strips in hallways and corridors, but only suitable for hallways and corridors (see Egemin Automation Inc/Bell & Howell Mailmobile). Now the time has come for individual point-to-point (desk to desk) robots.

Buy any of the four models of the SR4 Robot, depending on whether you will be adding your own triangulation and/or navigation computing gear (SR4-B, P or Pb), or using ours (SR4-N). Then, join the Office SIG of the Open Robot Network to compare notes with others working on the same challenges.

Partners Wanted

Smart Robots is actively seeking business and/or technical partners to join with us in the development and testing of applications, as well as in the introduction in various markets of the SR4 Office Robot. E-mail your interest to us at officerobot@smartrobots.com. We look forward to working with you.