W3C

- DRAFT -

Robots Breakout Session

29 Oct 2014

See also: IRC log

Attendees

Present
Debbie Dahl (W3C Invited Expert)
Kaz Ashimura (W3C)
Kazuaki Nimura (Fujitsu)
Gergely Ujvari (Hypothes.is)
Kristof Csillag (Hypothes.is)
Ryoya Kawai (AMEI)
Kirby Shabaga (Boeing)
Atsushi Kawata (Oro)
Yoshiaki Ohsumi (Panasonic)
Osamu Nakamura (W3C)
Yuma Inagaki
Metz Jonathan (W3C Invited Expert)
Ryuichi Matsukura (Fujitsu)
Tomoki Ogawa (Panasonic)
Han Jae Sung (Samsung)
Lori Glover (MIT/CSAIL)
Dowan Kim (Daliworks)
Tim Denney (Boeing)
Don Brutzman (Web3D)
Rumen Kynsakov (W3C Invited Expert)
Takuki Kamiya (Fujitsu)
Regrets
Chair
Debbie
Scribe
kaz

Contents


<kaz> scribenick: kaz

(list to be added later)

ddahl: robot is a complicated interaction system
... find scribe (=kaz)
... wiki
... (Robots)
... many form factors and capabilities
... purely social
... half human half mechanical
... (Then)
... speech interaction
... phone and desktop pc
... how do you interact with them?
... (Now)
... tablets, smart things for UI
... robots and appliances
... interact with robots
... showed these slides to introduce problems
... everybody is welcome to type into the IRC

(starting IRC on #robots channel)

ddahl: we can list use cases
... industrial robots, personal ones
... some of them move and others don't

<ddahl> scribenick: kaz_

<ddahl> roomba

<stakagi> transportation

ddahl: let's gather comments

<stakagi> avatar robots

ddahl: airplanes?
... robot cabin attendant
... interesting aspect
... could be relatively small

kaz: even like a bug?

ddahl: yes

kirby: smart luggage
... geolocation information is used
... you don't have to go to luggage claim

ddahl: how does a robot protect themselves
... what do you mean by "robots"?
... openstream provides pill dispenser robots
... medication pops up
... on the table
... mobility is not a requirement
... some level of autonomy is needed, though
... typically sensors at least camera
... security
... could be very small
... how does it protect itself?

<kirby> identity

tim: automotive field
... same cars underlying
... different size

ddahl: people around may be injured

tim: dangerous operation should be avoided

rumen: we get requirements?

ddahl: really web standards
... requirements for web technology for robots?
... application level
... core question is how do we interact with robots
... how do we use robots using a standard way
... probably you've seen some kind of robots
... robot interface is proprietary
... when we go to remote, that's also proprietary
... how can do it using standards?
... we can control robots using smartphones if it's done by standards

rumen: different bandwidth
... interested in seeing if there is common communication layer

<ddahl> scribenick: kaz

ddahl: different capabilities
... you want to be part of the ecosystem

metz: what if there is any open standard for chipset

ddahl: maybe need low-level standard
... HTTP over TCP/IP, etc.
... if you are not in know place, you have to use bluetooth, etc.
... maybe 3G or 4G would be better if it's in distance

metz: not just using bluetooth
... cars are also robots
... luggage could be remotely controlled
... by internet

ddahl: airport scenario
... using 3G connection to get it
... the luggage itself says "I'm here"
... transport question
... bluetooth and other protocols
... Zigbee
... very low power
... tiny little commands

metz: robotics should work in the real world
... allow a machine to talk other machines

ddahl: what you say is important

metz: protocol for robots

ddahl: something is needed as a high-level protocol for various low-level transport mechanisms
... one thing I might be thinking of is
... need an API to my personal robot, etc.
... inclined to do that
... "robot, play music"
... machine to machine commands
... multimodal group is working on MMI Architecture
... generic protocol for device control
... lifecycle events
... so generic
... can use for smartphones, vacuum cleaners
... not just start
... but start this and then start that
... describing capability of devices
... what can you do?
... devices reply, e.g., I can provide pills

kirby: discovery included?

ddahl: yes

tim: not library or programming language

ddahl: possible smart robot which tells stories
... smart luggage robot with self protection
... might want to have a supervisor
... security authorization
... also aspect of privacy
... camera and microphone might be very small
... there is privacy issues
... also accessibility issues
... maybe people can't see
... vocabulary for discovery capability
... e.g., published robot cabin attendant
... house cleaning robots
... describe discovery process
... this session is till 10:15

osamu-n: human robot interface
... machine to machine as well
... so many communication channels/technologies
... zigbee, bluetooth, wifi
... need a general variables on how to control robots

,.. we don't care of what kind of technology is really used for robots

scribe: very important viewpoint
... e.g., I order robots to play music
... we don't care about actual mechanism
... very important to control interaction

ddahl: use cases which require realtime connection?

osamu-n: realtime, safety, security, authentication
... want to know the capability of the communication channel

ddahl: car is a good example
... what is my capability?

(start to draw a picture on the flipchart)

scribe: high-level protocol and low-level protocol
... machine to machine communication?

osamu-n: yes
... robot to robot communication

ddahl: could be a smart building
... or two cars
... autonomous communication by smart things
... low-level communication and high-level one
... might be robot-specific protocol

metz: machine-to-machine use case is logistics
... different kind of products
... automated warehouse

kaz: maybe Amazon's storage

osamu-n: maybe there are 100 of robots there

<stakagi> swarm robotics via web

osamu-n: would avoid each other

taki: robot-to-robot or human-to-human as well
... asynchronous communication

<stakagi> For example, the communication with Rover on Mars has a very large delaying, and also its bit rate is also very low.

ddahl: definitely need offline use case as well

rumen: do we have technologists for this?

ddahl: we need standards

rumen: standards from IETF?
... constrained application protocol by IETF

osamu-n: recently working on HTTP2 as well
... compressed binary data is available

rumen: constrained application protocol is binary http

ddahl: websocket is also useful for realtime interaction

osamu-n: disconnected function is important

<brutzman> I posted an introduction message to the browserobo mail list at http://lists.w3.org/Archives/Public/public-browserobo/2014Oct/0000.html

osamu-n: important topic

ddahl: offline asynchronous

<stakagi> I think that there is already an offline web apps.

ddahl: would see what to do from standardization viewpoint

don: agree with all the comments
... shared vocabulary for robots will be central to this challenge

ddahl: requirements for standards?

(put requirements on the flipchart)

ddahl: offline operation

<brutzman> My university team's work on robot simulation is online at Autonomous Unmanned Vehicle (AUV) Workbench https://savage.nps.edu/AuvWorkbench

ddahl: scheduled operation
... realtime operation
... authorization
... discovery and vocabularies
... human to machine
... machine to machine
... privacy

kaz: one of the robots could be a virtual agent on pc instead of actual device

osamu-n: that's possible and one possible use case of "machine to machine communication"

ddahl: + virtual robot/agent (software) to robot
... + asynchronous operation

don: it's all the same thing if it's virtually simulated

ddahl: is that standardization for messaging?

don: yes

ddahl: + messaging standards
... human messaging and low-level messaging

taki: sometimes robot's capability reduces
... degradation of services
... the robot can provide some limited service, though

ddahl: + graceful degradation
... mmi wg will discuss discovery tomorrow afternoon
... at suite 1243

[ adjourned ]

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.128 (CVS log)
$Date: 2014/10/29 18:07:32 $