Collaborative, GPS-free Techniques for Localization in Miniature Robots

Localization within the environment is still a big challenge for robotic systems and especially for minimalist, insect-size robots used in swarm intelligent systems. Odometry is the simplest approach to this problem, and can be realized by keeping track of the speed of each wheel to calculate the robot’s relative displacement. Unfortunately, odometry is highly unreliable due to wheel slip and encoder noise.

In one of his projects, Collaborative, GPS-free Techniques for Localization in Miniature Robots, Nikolaus Correll is exploring a simple collaborative approach to reduce this error inspired by the way that flocking birds manage to return to the same nesting point from half way across the world. Taking simple flocking as a baseline, we will be able to further examine more complex collaborative algorithms.

Moreover, for analyzing collective behavior of mixed animal/robot societies, the SWIS group has developed SwisTrack, a tracking software allowing for tracking societies of a priori known dimension.

Relation to my thesis: Just being curious on the way other fields approach similar challenges on collaboration, space, context and failures/limitations. Collaborative localization system is stolen from the nature. It might have some value for the human world. Plus, they use tracking software to analyze collective behaviors. It has a common ground with the replay tool we use in Catchbob! to analyze group collaboration and location awareness.

The Internet of Things, Moving Beyond the Refrigerator

A Manifesto for Networked Objects — Cohabiting with Pigeons, Arphids and Aibos in the Internet of Things, thoughtfully written by Julian Bleecker, is an output from the workshop on Blogjects/Networked Things I participated to last month.

This manifesto engages us to think about the opportunities for enrolling Things into a thick and messy imbroglios of social interactions between humans and objects. With his blogjects, Julian brings humanist values (first-class citizenship, aiming at world 2.0, a more habitable world) to the RFID-enabled-refrigerator Internet of Things.

My perception of blogjects is that they are an application of the Internet of Things that I would preferably name a Web of Things. There are about generating content, meaning, and agency within the social web. Something that came out of the workshop (at least for Nikolaus Correll and I) was that blogjects hardly create an ecosystem. The exchanging and circulating (therefore I/O) information is not clear. I find the examples of very sophisticated robots (aibo) and animals (blogging pigeon) rather misleading. Primarily because I do not think that blogjects should necessarily have any “intelligence” (limited to a specific agency and actions). The “food for thoughts/meanings” emerge from flocks (interaction of simple individuals – swarm intelligence) rather than from animal/human-like behavioral individuals. It is the flock of pigeons that can interact with the Bay Area’s cars and drivers to make them aware of their own polution. Besides, would you like a pigeon to reach Robert Scoble‘s social status? Well, me neither…

A future Robert Scoble? (left). My vision of the Internet of Things while writing this…(right)

Relations to my thesis: In a pervasive world, objects interconnect and our perceptions and relations with them will change. I am interested in how this chaos can be managed. Features and scenarios around blogjects bring relevant clues to enter the post-refrigirator era of the Internet of Things.

Questions on Interactive Multimedia Systems

This morning I did a presentation about my perspective on interactive multimedia (yes, I hate that word too!) systems. Even done on the fly (assignment given the same morning), the exercise was very pleasant. I had to answer and discuss the four questions:

What kind of IMMs will we construct 5-10 years from now? 20-50 years?
What are the key enabling technologies?
What are the key enabling concepts?
What are your examples?
Please use your favorit artist/researcher/project to illustrate your arguments.

And here are the slides of my presentation.

Relation to my thesis: Among other topics, we got into the discussion on the abuse of the word “intelligent” (perception, cognitive, …). In the key enabling technologies, I only mention interface technologies and I do not talk about any approach (e.g. biology) to process information and render “intelligence” and emotions. This is mainly because I see these processes as peripheral and remotely believe in the smartness of single things. AI is facing the same issues with complexity. Google is not intelligent! My pragmatic approach applies to the path to the goal, not to the goal itself.

Pervasive Mobile Paradis

Le thème de cette semaine de Brèves Chronos Pervasive Mobile Paradis [pdf] est l’émergence du WiFi en tant que technologies sans-fil omniprésente. On y trouve notamment le podcast de Rafi Haladjan, patron d’Ozone, dans lequel il mentionne que l’enjeu est l’Internet (Rafi utilise le terme vague de “le réseau”) devienne naturel et fluide dans la mobilité de la même manière qu’il est inconcevable que l’on ne puisse pas téléphoner partout.

Ozone tenter de trouver un modèle commercial de masse pour l’informatique pervasive. Quelques points me dérangent dans la vision. Tout d’abord le connectivité fait constamment référence au milieu urbain. Les campagnes seront les lieux de déconnection et ou se trouvent les frontières? Puis la connectivité (le service) est vu comme un état binaire (j’ai ou j’ai pas, représenter par un % de couverture) faisant référence à l’accécibilité de l’Internet fixe. Le sans-fil a plutôt un nombre discret (voir même continue) d’états, sa couverture devant plus être représenté en bits/sec/m3. La fluidité ne pourra pas seulement être acquise technologiquement, mais également par un processus d’adaptation et d’appropriation de l’imparfait par les utilisateurs. Dernièrement, Rafi ne croit pas en la convergence de technologies sans-fils, mais a leur unification vers le WiFi en argumentant à raison que c’est la seule technologie grandement déployé ne demandant pas de l’acquisition licence. Par contre, la gestion d’une approche bottom-up (potentiellement chaotique) reste en suspends…

A ceci Brèves Chronos à dédié une section Connexion permanente et incertitudes. L’incertain fait son apparition dans les discussions non-académiques sur l’omniprésence et la mobilité.

Relation to my thesis: Ozone is a relevant bottom-up approach to pervasive computing for the masses. Uncertainty is now a topic appearing while talking about ubiquity and mobility. A sign of post-utopia and maturation of the field.

DragonForce

DragonForce, distributed by Drakontas is yet another awareness platform for emergency and security crews (similar to the swiss GeoGeny).The system allows commanders to send voice and text messages; communicate by drawing on clean slate or incident scene maps; and share images to the users carrying typical GPS enabled PDAs.

According to A PDA for police and soldiers lets them track colleagues using a digital map, Drakontas “had planned an aggressive marketing campaign starting next year for college campuses and public-safety agencies”. One of the current test exercise is the simulation of a high school hostage-rescue situation. My understanding it that they target campuses first because the wireless connectivity is controlled. Scaling this up to an uncontrolled environment such as a battlefield is another issues.

DragonForce will use VoIP instead of good old walkie-talkie systems. I am not sure where the improvements are. Moreover it is a very centralized systems that does not seem to support message propagation from one peer to the others.

A DragonForce client… hmmm or prototype.

Relation to my thesis: DragonForce is a real-world pervasive system that must rely on location and performant/reliable connectivity to communicate both voice and data. It would interesting to know how defense contractor Lockheed Martin Corp manages uncertainty in its similar systems supporting “battlefield intelligence”.

Worldmapper

Mark Newman, of the University of Michigan’s Department of Physics and Center for the Study of Complex Systems, has put together around 40 cartograms of the world covering such topics as population, tourist destinations, gross domestric product, energy consumption and child mortality. Data has been gathered from several different sources but mainly from UN organizations. Each map, comes up a small description which makes it more enjoyable. This is part of the Worldmapper Project.

World map of the GDP (Gross Domestic Product)

The cartograms are based on a variant of the algorithm produced by Newman and Michael Gastner diffusion-based method for producing density-equalizing maps.

Via the resourceful Cartography.

Relation to my thesis: I am interested in the ways to visualize information on maps. Cartograms could support the presentation to users the measured space sensed by ubicomp systems.

De Retour de Suisse

Data Communication and Mobility

A survey conducted at 3GSM by Tatara Systems found once again that the industry does not drink its own Kool Aid and actually use the services it promotes. For example only 24% of people surveyed at the show use their mobile phone to browse the internet. Nearly 93% of respondents use other networks beyond cellular mobile data with 43% citing the need for higher performance, 30% stating coverage issues, and 20% saying cellular mobile data was too expensive as the reasons for selecting other networks for sessions. Source: Stats & Research: 3GSM Visitors Low Users Of Mobile Data

Related: Notebook Owners Barely Show Nomadic Behavior

Relation to my thesis: Data communication while mobile is still marginal among practitioners. Lack of bandwidth, coverage, good prices or/and lack of real need?

Cityware – Urban Design and Pervasive Systems

The goal of Cityware is to develop theory, principles, tools and techniques for the design, implementation and evaluation of city-scale pervasive systems as integral facets of the urban landscape. It focuses on 2 major questions around space, infrastructure and its relationships with behaviors:

1. how do we design the space created by fusing electronically created interaction space with architecturally created physical space?
2. how do we provide interaction and interoperability that scales up to city-scale pervasive systems, while ensuring that they function appropriately and merge aesthetically with urban spaces, materials, forms and uses?

Three projects are currently defined

Radio City: the human behaviours associated with the presence of those radio signals.
Making space: explore the relationship between the spaces created by urban architecture and the interaction spaces created by artefacts such as digital devices
Where am I? What’s that? I’ll remember this: exploring opportunities that pervasive technologies offer to provide novel and effective support for way-finding, interpretation and recollection.

Relation to my thesis: I am involved in “ambient intelligence in the city” kind of project called ICING. In Cityware, I would be interested to know more about how pervasive systems scale up to city-scale while delivering appropirate experience to the users.

Intermittent and Planned Connectivity

The 3GSM Congress was the opportunity to do how mobile practitioners deal with mobility and connectivity. I witness some “Look I cannot talk to you, I am roaming…” and “Why are you calling me with a +41 number? It is going to cost me a fortune!”. Some 3GSM World connectivity blogged stories include Stuart Mudie’s Living la vida loca:

Well, firstly, I now know that typing on a phone is a real pain for anything other than the shortest of messages. [...] By the way, in case you’re wondering, I’m writing this post on my laptop. I have a neighbour with an unsecured wi-fi connection that I can sometimes catch if the atmospheric conditions are right.

and Russel Beattie’s Dial up… Wow, it still works:

[...] my “3 star” hotel in Barcelona is devoid of broadband, so after a few days of being disconnected, I was forced to root around my Y! laptop and find a corporate dialer, called the helpdesk back in Sunnyvale and asked for a password. [...] There is some intermittant WiFi at the 3GSM conference, but it’s not like I’m going to get a lot done in that chaos

Our vision of seamless integration and ubiquitous connectivity poorly supports real-world’s intermittent connectivity. The convergence via vertical and diagonal handovers of the telecom networks and wireless IP networks will only solve some connectivity problems and might create others. Connectivity is the trigger for data and voice communication. However the devices and applications used must deal with other constraints including bandwidth, coverage (autonomy of connection), latency, and last but not least the cost. Future successful mobile applications will need to plan connectivity by integrating these constraints in a subtle way (e.g. warning if gaming in a high latency area, synchronizing data only when entering a “free” wireless area). An excellent example of “planned connectivity” in the mobile industry was implemented by Shozu. Taken from Mobile Jones’ review of Shozu:

Data charges while roaming are brutal. I recently tried to upload some photos through Flickr and received an SMS from Shozu informing me that my transfer had been suspended, because I was roaming off my home network. Instructions for an override were provided or the option to do nothing, and the application would complete the transfer on my behalf when I was no longer roaming. Next, I received an SMS informing me that the transfer had been completed as I was no longer roaming. Now, that’s a pretty considerate thing to do, especially the part about managing this situation for me. True simplicity.

We manage intermittent connectivity on a regular basis. Like this out-of-office replay received this morning:

I will be traveling through March 3rd, however will be checking email during that time. If you have an urgent request, please contact xxx

Or, an idea I wrote down on my hand, because it was more convenient at that moment (while in a tramway):

Connecting has become implicit while now disconnecting has to be explicit.

Other inspiring planning of intermittent and planned connectivity come from some third world countries, like DakNet that provides mobile access points using motorbikes and buses to give Internet access to remote regions. They collect emails and provide information like a mail man would do. They call it cached WiFi intelligence.

This post is part of this week’s carnival of the mobilists XVI (sweet sixteen) hosted by gotomobile.