Sensewaves is a start-up at the origin of a Web platform for the integration of the connected and intelligent objects, which allows to analyze and to manage the information. This platform, Sweave, identifies the trends among the behavior collected by the connected objects, by transforming the raw data into significant information. A few weeks earlier, the Sensewaves team took away(gained) the EIR challenge Cisco for Sweave. We met one of the members of the founding team, Fivos Maniatakos to speak to us about this French start-up which rises.
History of Sensewaves
Sensewaves, it is a company of four people, among which three founders: Tommaso Bianco, Como Maestracci and Fivos Maniatakos, whom we met. Since a few months, a member was added to the team: Sophie Papachristopoulos, in charge of the marketing. Two consultants come to swell the ranks: Ken Forster and Philip Stockdale.The Sensewaves team with : Fivos Maniatakos, Sophie Papachristopoulos, Tommaso Bianco et Côme Maestracci.
The sound waves and the IoT
OBJETCONNECTE.COM : What is Sensewaves ?
Fivos Maniatakos : it is a platform of real time analysis of data coming from sensors, and dedicated to the industrial IoT. And we propose two things to companies with this platform: either we have tools easy to manage ( API) and thus the customer can integrate directly our analyses into the application, or one also has complete solutions, as it is the case with the load balancing.
OBJETCONNECTE.COM : What is the relationship between the IoT and the sound waves?
Fivos : In fact both are connected, because when we speak about IoT, it is in fact rather about data coming from sensors. And of chronological data, which we call scientifically time series of the temporal series). But in fact sound waves, (or ” digital audios “) are also signals of temporal series.
And it is where our students’ theses intervene in the history. My thesis, for example, concerned the improvisation of a computer, with musicians and musical instruments, in the very precise context where we had real musicians on stage, and where we wanted that a computer improvises with them. But the computer knows nothing in priori, because it knows nothing about their style. From then on the purpose it is to learn directly data without having of knowledge a priori. The system which was used then it is that a musician played maybe for a few minutes, and then the computer began playing with him. It is of the machine learning, we apply of machine learning, and after one in fact a representation
O.C : How took place this transition of the musical domain in that of the IoT?
Fivos : : At first our studies intended us all to be university professors, or researchers. But with our meeting in the IRCAM, we were fascinated by the idea of the creation of a start-up. It is as well at this moment that we saw that the IoT was exploding, and we said to ourselves that the actors of this market would badly need thorough analyses which we were capable of realizing in our domain. That it was in 2010.
Later we defended our theses and looked for a solution to make cross our knowledge in the field of the IoT. We worked on it in secret, and we prepared a first version of the platform, ready in October, 2014. Thanks to this platform we were able to win quite a lot of competition, while that makes only a few months that she works. Today we have a version alpha of the platform, and we are pursuing the process to realize contracts.
We realize POC (Proof of concept) to show the utility of the platform in diverse domains, as the industry for example. And we also work in pilot programs to validate our hypotheses, and move forward on the final version of Sweave.On réalise des POC (Proof of Concept) pour montrer l’utilité de la plateforme dans divers domaines, comme l’industrie par exemple. Et on travaille aussi dans des programmes pilotes pour valider nos hypothèses, et avancer sur la version finale de Sweave.
O.C : Why do you favor “time series”?
Fivos: it is a little bit particular. When we speak about IoT, we speak about ” chronological data ” and these have some properties which do not exist somewhere else. For example in a temporal series, the value which arrives just as we speak depends on values which arrived before. Then the method to analyze it has to take into account this property and differs in this point of tools collectively used in the context ” Big dated “. These data are raw, redundant, have a rather high transmit frequency. And thus are more difficult to understand and to analyze, especially in real time.
OBJETCONNECTE.COM : And it is from there that started the Sensewaves project?
Fivos Maniatakos : At the time of our theses, we were still in the part of analyze. It is after we passed in the action part: that is to know how we could adapt these analyses to the problem of a customer. And from this analysis, we were able to clear(release) predictions, analyses of trends, but also prescriptions, user’s profiling. The important later, it is to integrate that into the truth use compartment. For example, that of Load Balancing, which concerns smart cities, and district management
Concretely it is about small or bigger territories, which possess more or less autonomous installations of electricity, and where we are going to combine renewable resources, but while looking for new manners to consume the energy, as it is the case today with electric motor cars. There is a problem of the electricity networks which were much more stable before and which maintaining are not it any more, in particular because of all these various resources and of these various points of consumption. Then it is necessary to assure the stability of the network at any time, and to adapt it to the problems of Green energy, and of course make sure as well as the customer benefits from the best price, and it all the time from the consumption.
Today there are many actors who work on the elaboration of this kind of systems. What we propose we, they are tools of analysis, and deep analysis, to understand for example the behavior of the customers, make predictions, and assure the best articulation of the various analyzed parameters.
The future of Sensewaves in the IoT
OBJETCONNECTE.COM : How goes at present the market of the IoT according to you?
Fivos : The IoT is a universe which really moves for two years. And now we see that in this ecosystem there is great actors who manage various very specific problems. Some manage the connectivity, others are purely manufacturers of objects, and there is also the platform IoT which can allow the customers to create their applications. Then the question is to know how to position in this ecosystem there, how to find its place.
We designed the platform with the objective to make it complementary and interoperable with a majority of services proposed by the current actors in the IoT. It means that an actor – for example a designer of platforms IoT-can integrate our service into his product to improve his valuable proposal, because we supply analyses of time put into series that the other platforms do not have. And other types of actors take advantage of this solution, as those who are rather at the origin of the data, and who thus need analyses and tools, to look for example for recurring models of event, to have a deeper understanding on what takes place in their business.
OBJETCONNECTE.COM : You won the EIR challenge Cisco on May 29th of this year, what it implies (the only French company which won the challenge this year, 750 candidates and 5 winners at the end)?
It is the partnership of 6 months, and the partnership that delights us, because we were fascinated since the beginning at the idea of being able to work with a group such as Cisco, because it is the big giant of the IoT. Besides we are very close to the problem of Cisco, in particular in the idea that the universe IoT must be more organized. Today we have the feeling that a partnership such as this one opens our horizon, and which it is going to allow us to set up disruptive solutions in the IoT.
OBJETCONNECTE.COM : What are the difficulties when we are a start-up?
Fivos : it is necessary to dream, I think, it is necessary to see The Big Picture! ( The complete picture). And it is necessary to think that that takes time, that everything is not fixed from the beginning, and use our capacities of anticipation to know towards where things are going to go. And adapt itself non-stop to the ecosystem. It is also necessary to keep the eyes open, and to be attuned to solutions which emerge.
At the level of the difficulties, the one who was the biggest, it is that in 2011, when we started, people did not still know the term IoT. In fact at that time there we spent 60 % of our time to explain what it was and 40 % to explain our project. Thus it was necessary to be patient and to know how to explain simply what we made. Little by little, that changed, and now one does not any more need to waste time to explain what is the IoT.
The new difficulty is that people do not still see the added value of the IoT. And it is a domain on which came to be transplanted the crisis, as everywhere else. Thus people think more before investing, and are all the more suspicious when it is a question of investing(surrounding) on a quite new technology.
OBJETCONNECTE.COM : And in five years ?
We want to be leader on all which is analysis of data for the industry IoT. But also be able to support the start-ups which create an object, because even start-ups need sophisticated analyses. The objective would thus be to help them to launch more easily their ideas.
O.C : What is the projet you work on at present ?
We work on the heavy equipments in the industry which are very expensive, and cost even more when they break down. What we want to make, it is to supply to the customer the solution of monitoring that allows him to avoid a fatal breakdown. So we bring the possibility of assuring that the machine or the machines work at the most of their capacity, but also to avoid and plan early the breakdowns or the mechanical errors.