PIML stands at the intersection of data and natural laws. Where classic machine learning models learn purely from data, PIML adds a crucial layer: physical knowledge. Integrating these physical principles – think heat conduction, forces, flows — directly into the model creates a powerful hybrid that is smarter, more reliable, and much more efficient.

Why Physics-Informed Machine Learning makes a difference

For many companies, the challenge lies not in finding ideas but in translating them into robust products. Traditional development methods face their limits: slow iterations, endless testing campaigns, and a lack of scalability.

PIML breaks this pattern. By deploying a physics-informed model already during the initial design phase, you can make accurate predictions about a system’s behavior with limited data. So you don’t have to make thousands of measurements or do endless testing – the model already knows what is possible because it understands the laws of nature.

And that opens new doors.

• You can accurately estimate complex parameters such as temperature or pressure, even when direct measurement is impossible.
• You discover more quickly which design choices lead to failure probabilities or performance degradation.
• Your development cycle becomes shorter and more powerful: fewer iterations, fewer test failures, more learning capacity.
• You build systems that not only perform under ideal conditions but also remain robust under stress, variation, or unexpected situations.

Design faster, improve smarter

In practice, this means that companies like Verhaert do more than validate designs; we also take the time to truly understand them. Using PIML models enables you to predict component interactions, reduce error margins, and even optimize sensor placement – all before a physical prototype is built.

Innovation thus becomes not only faster but also bolder. Because when you know that your models give reliable predictions, you can take more risks in your design. You no longer have to test everything with brute force. You rely on insight.

Innovation is no longer a coincidence

With physics-informed models, product innovation takes on a new meaning. It’s no longer a guess but a thoughtful process. You’re not just developing a product – you’re building a system that understands how it will behave even before you build it.

For companies that want to be at the forefront of R&D and complex technology, now is the time to embrace PIML. Not just as a tool, but as an accelerator of vision.

The future demands robust products, designed in less time and with more certainty. Physics-Informed Machine Learning is the engine that makes this future possible.

What if your designs already followed the laws of physics? How would that change your innovation process? Let’s start the conversation on how PIML can transform your development process.

The post Meet the innovation accelerator of the future: physics-informed machine learning appeared first on Verhaert Masters in Innovation.

The post Meet the innovation accelerator of the future: physics-informed machine learning appeared first on Verhaert Masters in Innovation.

You can see a clear trend of moving from corded to cordless devices in product development. Creating battery-powered devices can be a challenge because of high-current drain and high-battery discharge cycles. In this webinar, Jan Buytaert Consultant PhysicsLab at Verhaert, discusses the successful development of an example device that is on the limit of the current battery technology, focusing on the development and testing strategy.

Watch the recording

View the presentation

 

The post High-current battery-powered handheld devices appeared first on Verhaert Masters in Innovation.

The post High-current battery-powered handheld devices appeared first on Verhaert Masters in Innovation.

Model-driven engineering isn’t the sexiest term to use in meetings, but drop “digital twin” and you’ve got everyone’s attention. Developing complex systems like robotic manipulators and AGVs uses empirical data gathering and theoretical modelling. Let’s explore the opportunities and pitfalls of virtual and hybrid prototyping techniques combining advanced system modelling and testing in real-life. This webinar was presented by Michiel Celis, Business Development Services – Product Innovation at Verhaert.

Watch the recording

View the presentation

 

The post Robotic innovation, the intersection of our digital & physical realm appeared first on Verhaert Masters in Innovation.

The post Robotic innovation, the intersection of our digital & physical realm appeared first on Verhaert Masters in Innovation.

The last decade we’ve seen these boundaries degrade to the point that CFD tools are applicable to everyday product design. The improvement of user interfaces, integration with CAD tools and automation of various modeling tasks have brought down the entry costs. There’s still a steep learning curve, however CFD has become an invaluable tool for more and more industries.

Let’s take a look at an example in the Fast Moving Consumer Goods business. Maybe you’re familiar with at-home beer dispensers or coffee vending machines with a multitude of products. Most of these appliances contain some kind of cooling system to cool down at least some of the contained products.

Home beer dispenser development

These appliances are governed by requirements of low cost, minimal material usage, space allocation constraints and ever more stringent power requirements. Squeezing out the last bit of performance has become a non-trivial task which can no longer be done with empirical formulas and hand calculations only.

Generic coolbox design (left) and CFD simulation of internal air circulation (right)

Here we show a generic coolbox design with internal air circulation which must cool down and keep cool 2 cardboard BIB containers. Hand calculations are fine to determine steady state conditions, but CFD can be used for so much more.

• Optimizing the placement of BIBs, air inlets and outlets for optimal airflow and cooling performance.
• Assessing the effect of cooling strategy on cool down time.
• Assessing the ideal control loop feedback for sensor locations.
• Calculating power requirements across various operating regimes and environmental conditions.
• Assessing the product temperature uniformity and prevention of product freezing conditions.
• Localizing cold spots which may cause condensation problems.

Building a single concept breadboard takes weeks and requires hundreds of hours designing, manufacturing and assembling. A CFD model however can provide answers in days. While a single cool down test takes hours, it can be simulated in minutes once a model has been created.

As such, CFD speeds up development cycles and delivers insights which are hard to obtain with empirical formulas and testing only.

Interested in finding out more on how Verhaert can apply simulations to your product designs or engineering problems? Get in touch.

Download the perspective

Looking for solutions to innovate?

Leave us your email and get in contact with Max van der Weyden, Manager Innovation Acceleration Services – Product Innovation, to help you with your innovation process.

The post Transient conjugate heat transfer simulation of a cooling solution appeared first on Verhaert Masters in Innovation.

The post Transient conjugate heat transfer simulation of a cooling solution appeared first on Verhaert Masters in Innovation.

About 21% of the European urban population is exposed to concentrations of PM10 (particles with a diameter of 10 μm or less) above the limit value established by the European Environmental Agency. 53% of the same urban population was exposed to concentrations exceeding the stricter WHO AQG value for PM10 in 2015. The percentage of the urban population exposed to levels above the WHO annual AQG (20 μg/m3) ranged between 50% and 92% in 2000-2015. The urban population’s exposure to levels above the more stringent WHO AQG for PM2.5 fluctuated between 82% and 97% in 2006-2015.

In the remainder of this document we’ll give examples focusing on one of the constituents being NO2, PM …

Measure what matters

If we really want to know how our air quality is, we need to start measuring it. Let’s take WHO AGQ threshold as objectives! Yet if we want change, we should consider the following objectives:

• Focus our community on what really matters in life.
• Allow us to measure our progress towards those goals.
• Enable large groups to work together in alignment (to start reducing NOx, CO2, PM levels not increasing it).
• Allow us to stretch to achieve things we wouldn’t have thought possible (Realize living conditions with air quality really below the WHO levels).

We cannot start acting without objectives on a global, regional, cities, district and even street level, as there’s a high urgency everyone starts acting in its own context at home or at work.

If we want to make cities more healthy and lively again, reducing the cost for society, we need to start measuring everywhere. And in order to be able to measure everywhere, the measurements or their measurement devices should be cheap enough. This perspective covers a set of technologies that Verhaert masters and which we would like to share.

Optical measurement technologies

Lots of possible measurement technologies do exist like ODS, laser diffraction, scattering, ring-cavity … Although Verhaert’s OpticsLab has expertise in all of them, we’ll discuss briefly the 3 most common of them:

1. LIDAR
2. Scattering
3. Satellite EO data

The following measurement principles must operate in a closed cavity to avoid disturbance from environmental light, thus making the system mechanically somewhat more complicated. However the scattering principle remains a cheap solution from which many instances are already available on the market.

LIDAR

LIDAR technology has become more advanced with the development of laser technology to include the following remote sensing techniques:

• Elastic-backscatter LIDAR
• Coherent, Raman or Doppler LIDAR

Elastic LIDAR measures the changes in magnitude of the reflected light, while coherent and Raman LIDAR provide information on changes in the wavelength of the reflected light. Differential Absorption LIDAR (DIAL) was developed primarily for the spatial measurement. Typically it contains a spectrally separating optical component, directed into the detector.

The DIAL measurement is achieved by the direct impingement of the laser beam on these materials and its subsequent reflection and scattering. All LIDAR applications use fairly high-end laser and optics. They’re typically designed to measure over long distances and to achieve high precision measurements as requested by Earth Observation Laboratories or Meteorological Institutes. Technological advancements in this field are still made today for specific performance challenges. Chinese investigators developed in 2017 a lower cost DIAL system with a detection sensitivity of ± 0.9 ppbv (parts per billion by volume) at 95% confidence level in the region of 0.3-1 km.

Scattering

Laser diffraction or scattering is often used to count particles and determine their size and concentration in the air. When a laser beam strikes a particle, the beam’s light is scattered in all directions. A light detector (or ‘photometer detector’) measures all of this scattered light. Larger particles scatter the light more than smaller particles. An algorithm determines how many particles are in the sample, and how big they were. These sensors are able to measure concentration of near microscopic particles ranging from 0.5μm (and sometimes a bit lower) up to 10μm.

An analysis of the histogram plots shows that the pulse height (amplitude) has a direct relationship to the size of the particle. The sensor outputs larger pulses (higher ADC Code bins) for larger particles. This result could be expected considering that larger particles will scatter more light compared to smaller particles. The figure below shows the histogram of ADC Codes for Smoke Particles and dust.

Sensor detection pulse’s amplitude is related to particle size and sensor detection spacing between 2 pulses (i.e. frequency) is related to particle concentration. The software algorithm estimates the concentration of particles flowing through the sensor by measuring the pulse amplitude and spacing between pulses detected. As the PM2.5 definition includes only particles with a diameter of 2.5 μm or less and as particles above 2.5 μm will generate a pulse that meets a higher amplitude, a threshold can be set to identify particles smaller than 2.5 μm. These kind of systems can be built with low cost photodiodes, laser diodes, filters, low cost lenses and appropriate processing electronics, but will only be accurate through calibration.

EO data – Satellite based measurement

Why bother about you having a proper measurement device when EO satellites (Earth Observation) can offer a free of charge solution. Different scientific instruments aboard NASA and ESA’s EO satellites assess the Aerosol Optical Depth (AOD). This is a measure of the degree to which aerosol particles prevent the transmission of light either through absorption or scattering. Several studies have developed algorithms and models to relate the AOD measures to ground-based measurements of particulate matter and gases like NO2, CO, O3, SO2 … Of course, relating these atmospheric column measurements to ground-level measurements is tricky and depends on the vertical structure, composition, size, distribution and water content of the atmospheric aerosol.

NASA has several EO satellites like AURA, AQUA, CALIPSON, LANDSAT and TERRA. The first of a kind Moderate Resolution Imaging Spectro-Radiometer (MODIS) mounted in the TERRA and AQUA-satellites measures spectral radiance from 412 till 14.200 nm enabling a lot of spectral based gas concentration measurements. Results indicate that the mean differences between PM2.5 reported by ground monitors and PM2.5 calculated from the satellite overpass times during cloud-free conditions are less than ± 2.5 microgram (-3), although this value varies by season and location. TERRA was launched already in 1999, and AQUA in 2002. In polar orbit, approximately 700 km above the Earth, MODIS views a swath of approximately 2.300 km resulting in near daily global coverage of Earth’s land, ocean and atmosphere system. Air pollution from Aerosol products are reported at 10 km resolution. NASA’s EO details can be found on search.earthdata.nasa.gov.

But also ESA has a large EO satellite track record: PROBA-V (developed, built and commissioned by Verhaert in 2013) and Sentinel 1-3. ESA recently (October 2017) launched its Sentinel-5 Precursor – also known as Sentinel-5P – the first Copernicus mission dedicated to monitoring our atmosphere. It has been built to map a multitude of trace gases such as nitrogen dioxide, carbon monoxide and methane that affect air quality.

Carrying the Tropomi instrument, the most advanced of its kind, Sentinel-5P maps the entire planet every 24 hours. What sets Tropomi apart is that it measures in the ultraviolet and visible (270-500 nm), near-infrared (675-775 nm) and shortwave infrared (2.305-2.385 nm) spectral bands. This means that a wide range of pollutants such as NO2, O3, CH2O, SO2, CH4 and CO can be imaged more accurately than ever before. With a resolution as high as 7 x 3.5 km, it has the potential to detect air pollution with a district precision. Images of the central part of the Flemish region for a working versus weekend day are given below.

Of course using historical data, air pollution forecasting models exist f.e. ENSEMBLE model. Now given for PM2.5 for Europe.

With all these great technologies it becomes possible to monitor what’s important to us. Yet it’s only through building new specific applications addressing needs for users in the streets, applications for districts, managing cities and regions that one shall capture value in through air quality monitoring.

Acting collectively and stretching together to achieve reductions
in air pollution.

If we can get large groups working on finding viable and valuable products, we can switch from measuring to reducing NOx, CO2 and PM levels, not increasing them. If we act collectively, we still can limit the pay back of air pollution.

Performance criteria and test procedures for certified Automated Measuring Systems (AMS) for ambient air quality for gases and PM are defined in:

• EN 14211:2012 for NOX
• EN 14212:2012 for SO2
• EN 14625:2012 for O3
• EN 14626:2012 for CO
• EN 14662-3:2005 for C6H6
• EN 12341:2014 for PM10
• EN 14907:2005 for PM2.5
• EN 16450: 2017 for PM10 and PM2.5

Download the perspective

The post Measuring air pollution appeared first on Verhaert Masters in Innovation.

The post Measuring air pollution appeared first on Verhaert Masters in Innovation.

Sensors and technological trends go hand in hand. Sensors and metrology are key in driving technological innovation of today. Industry 4.0, IoT, Smart Mobility, autonomous driving, Smart Energy, Smart City, condition monitoring or industrial automation, in these industries sensors are becoming ubiquitous and connected.

But what is a sensor anyway?

Back then a sensor was just a component producing an electrical signal from a physical value. Easy no? Nowadays, even simple sensors have evolved into intelligent systems with self-calibration, self-control, digitization, signal processing and digital interfaces or even on-board data logging. Today, digital parameter and values indicating an anomaly are transmitted to remote surveillance centers. The next big thing is a smart autonomous sensor system. Here, next to the registration of one or more measured variables, the entire signal processing and power provisioning or even energy harvesting is incorporated in one housing.

Typically, these smart sensors include a microprocessor with communication interfaces that provide connectivity to other sensors (sensor mesh network) or to a central system (locally or in the cloud). This way, the sensor is able to identify and extract complex, application-specific, useful information from one or more physical property data streams. Thus, sophisticated tasks can be achieved by the sensor without an external computer. And with this, the demand for smaller, decentralized, more reliable, cheaper and more flexible systems can be achieved. But these requirements are never entirely fulfilled and there are future sensor-generations to come which are even better, smaller, smarter, cheaper and more autonomous.

Smart sensors widen the value chain

With smart sensors, we mean the intelligent combination of sensor data with other information. Be it the localization through GPS or WLAN or model-based knowledge about the process, so that sensors, along with algorithms, can already decide in the field, which incident needs to be reported. In parallel, the intelligent combination of various sensor data has increased importance. Smart autonomous sensor systems allow the users of these systems to widen their own services and offerings through to completely new offerings. In the future the co-operation within the entire supply chain becomes important. Next to smart sensors, new business models for smart services will become possible by utilizing intelligently data of products and manufacturing.

Envisioning new opportunities

With smart autonomous sensor systems, along with IoT, AI and cloud-based solutions new opportunities arise that only work when the supply chains or industries are connected. It becomes a matter of seeing the opportunities. A recently founded, inspiring place may be the IoT-Innovation space in Munich. This is a cooperation between Deutsche Telekom and IBM. Here it is demonstrated, how applications like smart parking, intelligent waste- and water-management, air quality control and predictive maintenance work. One example is sensor-generated air quality data that are aggregated and combined with weather-prediction-data and traffic-data. With this, users in the city council get insights about air quality and its dependencies and end users can get information about air quality via an app.

At Verhaert Masters in Innovation, one of our key competencies is to create smart technical solutions which fulfill the customer needs in an optimized way, which is not available off-the-shelf. To achieve this, we combine multiple disciplines like electronic design, software engineering, algorithm development, simulation, mechanical design and prototyping. Whilst developing these technical solutions, the user and business aspects are taken into account and new value chains are created.

Here are 3 examples of smart autonomous sensors.

1. Turbo speed sensor for TE Connectivity

Cars are everywhere around us, growing more performant, efficient and even autonomous. As an automotive part supplier, TE Connectivity is constantly reinventing its sensor portfolio to meet these automotive requirements. One of its new sensors is the turbo speed sensor. The turbo(charger) is a device used to recuperate the remaining energy in car exhaust gases and use this energy to pressurize the engine inlet air. This action is performed using a free-running turbine wheel, of which a maximum speed is not to be exceeded to prevent damage. To find and regulate to the optimum between inlet air pressure, wheel speed and engine performance, the speed of the turbo wheel needs to be measured. In regular cars however, no turbo speed sensor is available, and a high amount of margin is used to prevent this, as no assumptions can be made on the location (and especially the air pressure) of the car, leading to a suboptimal engine performance. The goal of this project is to improve the engine performance of all regular cars by making turbo speed sensors generally available. The turbo speed sensor performs this measurement and transfers the result to a car controller unit.

The sensor needs to meet the following requirements:

• The sensor needs to work from 0 rpm to 320 krpm of an aluminum or titanium turbo wheel. This corresponds to the turbo wheel speed of all existing cars. Especially the low rpm range would be an advantage when compared to the competition.
• As the automotive industry has high safety and reliability requirements, and the turbocharger environment is hostile to electronics, the temperature requirements were stringent and the sensor needs to comply with OBDII self-diagnostics. Especially the high temperature range mainly limits the component portfolio to passives, transistors and simple micro-controllers.
• Due to very high volumes, the cost and mass manufacturability were also very important factors.

Verhaert used its integrated product design approach (inverse the learning curve) to obtain a sensor solution meeting these challenging product requirements, with limited design effort. This trajectory started with a short study of the turbocharger environment and a sensor technology scouting. This scouting exercise resulted in a shortlist of rotational speed sensor principles, ranked based on earlier experience, applicability and technology readiness level. Examples are: inductive and mutual coupling, capacitive, acoustic, variable reluctance… This list was further reduced by performing simple tests ‘fail fast, learn fast’, indicating the validity of each of these measurement principles in this application. In the end, only the capacitive and inductive/mutual coupling approach remained.

For each of these, multiple implementations were made, consisting of a sensor head and driving electronics. Circuit- and PCB-design were used to optimize each electronics design, where 3D EM-field simulations were performed to design the sensor head. Each of these implementations was functionally tested in combination with a turbocharger, and was graded for each of the requirements.

This design approach resulted in a patented mutual coupling sensor implementation, which was not only highly effective but also lean, low power and low cost, as, after optimization, only 2 transistors, a few passives and a single micro-controller were used. As the sensor head needs two angled coils, which has a higher manufacturing cost compared to an inductive sensor head, the cost advantage is slightly reduced. Still, the high output signal quality, leading to low post processing requirements, reliability and versatility makes this sensor design able to cope with tomorrow’s challenges, both in automotive as well as in other industries.

2. Smart corrosion sensor for Resus

Corrosion is one of the main causes for failing heating systems. Constant monitoring and early detection of corrosion inside the heating pipes can give vital information on the system health, prevent system breakdown and avoid extensive repair costs.

Verhaert developed and optimized a sensor principle that’s able to measure the corrosion level. The sensor is mounted into the water flow via a T-coupler along the heating pipe circuit. The tip of the sensor contains a small, thin steel plate (approximately 100µm thick). The thickness of this plate is monitored through Eddy currents. The water in the heating pipe will corrode the thin plate, hence changing its thickness. By monitoring the change in thickness of the thin plate, a reliable prediction of the corrosion status of the whole water piping circuit can be made. This has been proven by extensive testing under various conditions. An advanced signal processing algorithm allows for a threshold to be set to give an early warning when corrosion level becomes a concern. This alarm triggers a contact or can be sent to a monitoring system.

Verhaert designed a unique, stable measurement system with low power consumption and long-time stability. Furthermore, we developed the mechanical parts and designed a hermetic seal that withstands the large temperature variations under higher pressure.

3. BOXX smart meter for Engie

As a leading energy provider, Engie is looking to offer their customers more added value via services. BOXX is a platform that allows end users to monitor and manage their energy consumption (gas, electricity, water). BOXX can be considered as an intelligent thermostat, but since it is a connected product, it can offer much more. In order for BOXX to work, it needs primary data input concerning energy consumption. For measuring electricity, Engie involved Verhaert to develop a sensor that could be retrofitted to the existing installed base of Ferraris meters (meter with turning disk). This type of meter is still the de-facto standard until it will be gradually replaced by intelligent meters in the coming 5 to 10 years. The list of requirements was very challenging:

• work with all types of Ferraris meters present in the field (>30 different types)
• non-invasive mounting on meter (meter cabinets are sealed)
• easy and quick install (with regard to current clamps)
• removable
• detect power consumption and production (for PV equipped home installations)
• very low cost
• work under all circumstances (dark, sunlight, artificial light)
• self-calibrating

Verhaert came up with a solution that is a smart autonomous sensor combination of optical technology (custom developed lens system, carefully selected led light source and modulation), analog signal interface and advanced signal processing. By tackling the job-to-be-done on different fronts concurrently, Verhaert was able to meet all requirements. The energy sensor is currently on the market with a steadily growing installed base. Customer feedback confirms that the sensor can be installed very easily and works reliable in almost every situation.

Download the perspective

<![endif]–>

The post Smart autonomous sensor systems appeared first on Verhaert Masters in Innovation.

The post Smart autonomous sensor systems appeared first on Verhaert Masters in Innovation.

Technology horizon: prepare the future of products

Last year we turned numerous technologies into demonstrators for Internet of Things (IoT) applications to evaluate its possibilities. These IoT demonstrators focus on the digital transformation of your products and services, creating new functionalities and user interactions.

Creating collective intelligence. We used multiple car sensors from many vehicles in combination with a data feed of weather forecasts. It enabled us to predict real-time weather road conditions by using self learning probabilistic models with spatial and temporal distributions. The result? Location based augmented weather maps, created in real-time. The maps are available in every car, at every location. Based on the local weather information, they can take safety actions: reduce speed, warn for slippery, switch fog lights, change navigation etc. We see interesting applications in, among other things, agricultural vehicles and heavy duty construction machinery equipped with sensors for crop presence, crop quality, excavation duty or precision information, payloads weight, etc.

Creating highly aspheric optics for low cost sensors. We prototyped custom designed wafer optics in an optical lens stack design with highly aspheric surfaces. This production technique is typically used to create cost effective optics for mobile phones, but then the optics has been designed for other applications. We realized a technology demonstrator with a high resolution (ca. 20 μm), large field of view (FOV diagram = ca. 90°), large depth of field (DOF = 3-14 mm) and low distortion (below 5%). This deadlock situation needed to be resolved: the high resolution was in contradiction with the large depth of field (since the spatial resolution improves with increasing numerical aperture while the DOF degrades with increasing numerical aperture). The desired characteristics of the optical system were inconsistent with each other from an imaging science point of view. Therefore they were difficult to satisfy within one optical system. But we realized a demonstrator coming real close to what’s possible.

We became MFi (Made for iPhone/iPod/iPad) certified to develop hard- and software meeting Apple specific design rules in order to communicate with Apple devices. As a result Verhaert has not only access to the technical specifications and sample quantities of licensed components, but also to Apple’s development tools and technical assistance to support your product development projects.

Amazon Alexa is a real innovation enabler. We noticed a tremendous increase in use-of-voice-controlled assistants as the offered convenience has already faced down conservatism. Alexa is a new layer that makes your device machine UI (or part of it) not only obsolete but also smarter since Alexa takes over your operator tasks. It serves as one generic interface controlling a distributed set of connected machines. Alexa doesn’t only manage your schedule, arrival and departure time and databases, but it also takes appropriate actions. Verhaert used Alexa on a custom PCB hardware for a new home device. We configured Alexa in order to manage new interactions different from today’s MMI (Multi Media Interface) models, creating new possibilities depending on the dialogue, integrating several other devices in one conversation and controlling them all.

What else is coming up later this year?

Realizing a first LTE-M and NB-IOT technology platform to connect any product. LTE-M (3GPP Release 13) might set a new standard, impacting LoRa and SigFox networks heavily. Providers are busy upgrading their networks to the latest technology and soon will guarantee 100% coverage worldwide. With a fall back to the precursor data-channels of the GSM-network, it will enable the world’s first mobile Smart Autonomous Sensor Systems. We’re ready to deliver a module and custom antenna design, our knowledge will serve any request.

Mid-air Haptic Ultrasound interface. New interfaces define the winners in a lot of applications. New use case design leads to a better way of interaction accommodating in the context of f.e. driving a car or surgical handling in an operation room. Voice User Interfaces will help, but still leave the user with a nostalgic feeling for feedback. Therefore Verhaert built a demonstrator of a mid-air haptic interface to deliver the next generation of user interaction. This new haptic interface feels like you touch a button in the air, just like real.

Artificial Intelligence (AI) in healthcare to assess blood pressure from photoplethysmogram (PPG) signal. From a noisy PPG we’ll able to estimate blood pressure to quantify hypertension levels. For this we need a lot of datasets (expert labelled), which are much easier when they are available in databases than they must be measured newly on patients. We evaluated different online sources (Mimic.physionet.org, Archive.ics.uci.edu, Catalog.data.gov, Capnobase.org, Kaggle.com, etc.) Although none were a perfect fit for the project goal, it allowed us to start prototyping the deep learning algorithms much faster and use transfer learning techniques to solve the problem.

Innovation horizon: add features to capture the value

Research over the past year identified that in order to make digital transformation a success, you have to connect technologies with the creation of new businesses or the disruption of your existing business. Just don’t stop with connecting the sensors or device to the cloud. The technologist’s mistake is often to stop there. Therefore we added this second horizon to help you defining features on how buyers and users can capture the value of it.

How to design new value into a product? Your business might be in the situation from eroding. Often it’s not the cost of the product that’s the problem, but the value that it’s not well conceived. Features that don’t carry user appreciation don’t contribute to the willingness to pay. What are the alternative options at that moment? Instead of choosing for a design to cost or modularization project, you might consider a value design exercise.

• Redesign your product to deliver a new value/cost ratio and extend the current product life cycle. F.e. revisit the combination of features (incl. new features) that customers really value and get rid of the unvalued ones. This off course leads to developing a new product which is combinable with a drastic design to cost exercise.
• Extend your current product and build an ‘add-on’ (MVVP) with a separate value yet creating a combined new value proposition on top of the existing one. When you aim at a digital interface for your product, you might add regularly new software features to keep adding value.

Did you consider the ‘whole product’? Adding a layer such as digital is a unique way to enhance your product with functionalities or reach out to a new target group. You could even allow third parties to start developing and offering their proper layer. But the aim of these layers should be to address multiple applications (one per target group) and the needs of the early and late majority. By identifying additional features added on top of the initial MVP, one aims at crossing the chasm and hence reaching a larger audience and target group.

Design for behavioral change. The world is changing rapidly. Current products on the market aren’t helping users to adapt and change their behavior to the world around us. If you want your products to help your users and customers with this change, you’d better implement nudging features into product concepts. People can be very reluctant to change practice, even if there’s evidence saying that they should. However some of the primary barriers to adopt technology are time, usefulness and complexity.

Let’s resume a clear objective we could and should try to achieve with this focus on implementing nudging features.

• A company needs to focus on highly appreciated features to create highest user adoption as possible. You can differentiate by being the first in the market that addresses latent user needs, but are your customers also ready to change their current practice? In order to accommodate your customer to the new use, those new functions, you’d better help them to change. Design for behavior will be a promising way to address societal challenges, so keep tuned. For technological enterprises changing consumer behavior is number 4 priority according to a PWC market research!

2018 will probably be remembered as the year when data driven business clearly emerged beyond the stock performance of the incumbent companies. When looking to the Plattform-Index it’s clear that platform business is an interesting business model outperforming the rest of the market. Transforming the current business model into a digital version and developing a platform to attain maximal customer lifetime value should be a board priority. But in order to do so, internal processes, services and products need to be transformed.

Combine horizons: achieve more

IoT is there to stay, so you should give it a high priority in your innovation strategy. Developing products today, doesn’t mean products of today, but products of the future.

Rule 1: It’s all about envisaging ‘use’ for the digital era to come. Deploying the newest technologies to develop new functions and evaluate how they make products more useful or desirable, will help you to create options of tomorrow.

Rule 2: Developing products without a clearer strategic goal of what one wants to achieve, is predominant to get focus, secure ambitions and well consider risks.

Download the perspective

The post Unleashing the power of IoT appeared first on Verhaert Masters in Innovation.

The post Unleashing the power of IoT appeared first on Verhaert Masters in Innovation.