Welcome to Signal Cruncher!
Embedded Machine Learning for IOT
We bring local and safe AI to you
We offer local Artificial Intelligence (AI) that runs on 100€ hardware to maximize user benefit.
While Internet Giants use central AI on 100 mio. € cloud server farms to influence people.
With almost 80 years of AI Experience
Our collective experience helps you to find proven solutions to your specific challenges,
unlike so many firms currently, that follow the hype without real understanding of AI.
Helping You to earn money from IoT
Smart Buildings understand their users and capture the mass market. Smart Energy can reduce primary energy cost by 30%.
Smart Factories increase margins by 3 to 10 percentage points.
Helping you to earn money from IoT
Helping smart home system suppliers to reach the mass market by ease-of-use and ultimately offering an artificial concierge that helps the user to perfect his/her home.
Heating and cooling system providers and utilities are expanding their portfolio towards complex solutions for their B2C or B2B clients. AI is the key to reduce their energy consumption even further.
Global manufacturing markets have reached a level of complexity and uncertainty, that requires AI to further improve supply chain, procurement, production yield and profit
We are ready for Smart Building
– Programming is for a few, learning is for everyone –
Value for consumers:
Today – “Imitating user’s behavior”
Instead of having to program home atomization by rule-engines, AI can just learn from inhabitants how lights, shutters, blinds, alarms and much more is used and is able to repeat these patterns upon request.
Tomorrow – “Satisfying user’s needs”
When basic user needs (e.g. security, saving money or comfort) are understood, AI can continuously optimize home automation in dialogue with the users.
Value for smart home providers:
Today – “Access to the mass market”
Since only a small minority of the population is able and willing to program rule-engines, today Smart Home is a small niche. Self-Learning systems are for everyone.
Tomorrow – “Automated targeted upselling”
As “Trusted Advisor”, the AI not only inspires better use of existing components but also suggests new components and applications to extend the smart home to continuously increase user satisfaction.
We are ready for Smart Energy
– AI helps to consume only want’s really needed –
Value for consumers and business:
Today – “Reduce comsumption by 30%”
AI learns thermodynamics of the building in changing weather as well as the users’ comfort-limits of each room’s temperature over time to control furnace output, thus minimizing consumption of gas, oil or electricity.
Tomorrow – “Minimize total cost of prosumers”
Optimizing overall energy cost in complex systems of decentralized production (e.g. solar, bi-generation), controllable consumption (e.g. cold storage, heating), storage (batteries and heat buffers) and bi-directional grid usage – in this kind complexity AI offers vastly better results than traditional control algorithms.
Value for smart energy solution providers:
Today – “Clear differentiation”
Manufacturers of heating and cooling systems as well as utilities can secure and grow market share by offering highly efficient systems. In addition such modern optimization capability helps to establish them as a provider of solutions beyond hardware.
Tomorrow – “attractive new business models”
As continuous optimization in complex decentralized energy systems allows for significant cost reduction, Contracting („Renting“ the system incl. service) becomes more attractive than purchasing. Thus allowing the provider continuous revenues with high customer loyalty/lock-in.
Future growth in Smart Factory
– Taking it to the next level with AI –
Value in demand forecasting:
As demand becomes ever more volatile and order lead times for components get longer, AI can help by continuously learning from market patterns to create up to 50% better forecasts.
Value in procurement:
As an ever growing number global suppliers become more sophisticated in optimizing price, quality and availability for themself, AI offers optimization for manufacturers far beyond economies scale – while cutting cost by up to 39%.
Value in sales of services:
With increasing competition and pressure on prices, generating extra revenues by separately selling product-related services is getting in focus of many manufacturers. AI can boost EBIT by up to 13% by finding the best customers for new service offerings.
Value in production:
The steady increase in complexity and the need for ever more flexibility has reached the limit of humans and traditional algorithms. AI can help to reduce faults to 1,2 ppm, improve availability by up to 40% and improve overall production yield by up to 10%.
Value in supply chain:
Leaders in AI-application reduce their inventory by up to 50%, related handling cost by up to 10%, transportation cost by up to 12% and supply chain administrative cost by up to 40%.
Value in product development:
As designs become more complex and development cycles must be ever shorter, AI simulates and optimizes electromagnetic compatibility and manufacturability, thus reducing time to market from years to weeks.
How we optimize building automation
Minimal hardware requirements, that allow cheap and local AI applications.
Reinforcement learning requires no storage of transactional data (increased data protection).
Our best stochastic algorithms require up to 1.000x less processing power than deep neural nets.
Pilots show, that trust-based intimate interaction of humans and AI with contextual knowledge yields great predictions on ”a sample of 1“ instead of millions.
Proven library “XONBOT”
- Bayes‘ networks and regression
- Regression trees
- Multivariate splines
- Deep neuronal networks
- Factorization machine
- Support vector regression sparse grids (only available in XONBOT)
- Real-time learning in milliseconds
- No storage of transaction data
- Powerful constraint handling reduces complexity for the stochastic algorithms
- Optimization of long term interaction instead of viewing every decision seperately
We make it happen
Over 80 years of experience to help you to profit from AI.
Dr. Michael Thess
30 years of applied mathematics, significant AI publications, e.g. on sparse grids. Serial founder, among others co-founder of prudsys AG, a leader in AI for European eCommerce.
Founder for customer centric web services with more than 20 years of experience. She gained extensive experience in the implementation of technology-based concepts.
Four years ago, the proprietary Java library “XONBOT” was launched. He is fast, lean and strong in continuous learning with predictive optimzation. He keeps data local and safe.
Why you should start now
Signal Cruncher und Fraunhofer IPK
Teaching Machines how to Save Electricity.
Signal Cruncher and Fraunhofer IPK explore how manufacturing companies can reduce their energy consumption with the aid of machine learning.
Signal Cruncher is part of the ReLkat funding project
Signal Cruncher is part of the X-Kanban funding project
Signal Cruncher is Part of the PwC Next Level Scale Ecosystem
The scale program brings startups, corporates and strategic investors together and specifically addresses the challenges of all players.
Experience XONBOT in action
Watch how XONBOT learns to control the Smart Home, by repeating how lamps are switched on/off, which increases user comfort and can be used to deter burglars, by pretending someone is home
Signal Cruncher: AI for Smart Living solutions.
Signal Cruncher and prudsys