Welcome to the
Project Configurator
Here you will find our list of services and competences for research and development projects. Select several project modules, technologies or other details from our six areas of expertise to put together your project with embedded AI.
Hinweise:
- Decide for yourself whether you want to make a detailed or only a rough selection. Add your wishes in the corresponding input fields.
- You will then receive your configuration as a report for download (anonymised) or send us your configuration as an enquiry.
-
ProjectSelect project form, area of application and project management
-
SignalsWhich signals, sensors and signal processing algorithms describe your project?
-
AI AlgorithmsDo you need recognition processes, machine learning, AI or other algorithms?
-
SoftwareWhat are your software development requirements for your embedded AI project?
-
EmbeddedDo you want your embedded AI to run on an electronic board? We design your embedded system according to your wishes.
-
ConnectivityYour embedded AI doesn't work alone? Choose connectivity options with other systems.
Project
The project or project work is the basis for professional research and development work, series production or product design. In this section, you can provide more detailed information on the environment of your project, the area of application or the project form. You can also select the type of project management or find out more about it. You can also check whether our methods are already being used in your company so that we can adapt to your approach.
![]()
Project description
Please provide us with a verbal description of your project so that we can get a good idea and prepare for a collaboration.
Industry
In which industry or area of application does your project take place?
Project type
Product development typically goes through several phases that cover the entire product development cycle. Which phase of this cycle are you interested in or where can we provide support?
Project Management
Embedded AI offers you professional project management (PM) to ensure the success of your project. We have various methods at our disposal, from which you can select some aspects below. If you have no preferences, we will suggest a project management approach.
Project management procedure models
The choice of project management procedure model is based on the project type and the customer's external requirements.
More precise selection (optional):
Projekt Management Tooling
Embedded AI works with the project management tools listed below as standard. Select a tool of your choice for your project.
More precise selection (optional):
Requirement Engineering
How is the requirements engineering carried out? If desired, we can develop the technical requirements for your project idea. We take over the collection, documentation and management of the requirements.
More precise selection (optional):
Defect Management
Defect management is a systematic process for identifying, documenting, analysing, prioritising and eliminating errors or defects in a product or system. Do you have specific defect management requirements for your project?
More precise selection (optional):
Quality management (norms, standards, certifications)
Project quality management refers to the systematic monitoring and assurance of the quality of all project-related processes and results. It includes the planning, control and improvement of quality aspects to ensure that the project fulfils the defined quality standards. Certificates attest to certain industry-related standards for quality management or quality. Is one of the following standards or certificates required for a project?
More precise selection (optional):
Signals
Embedded AI projects often deal with the utilisation, evaluation and recognition of information in signals. Signals transmit information from an information source (e.g. a face) via a signal transmission medium (e.g. infrared light) to a sensor (e.g. infrared camera), which converts the information into a machine-processable form (e.g. digital image). A subsequent digital signal processing algorithm can transform and modify these signals or use them to extract information.
In this section, you can find out about available signal forms, sensors and signal processing algorithms and make the appropriate selection for your project.
![]()
Physical signal shape
Embedded AI can generate solutions with all signal types. If your project is to process input signals, please select the type of signal below. If your signal type is not listed, please contact us.
Acoustic
An acoustic signal is sound in the human audible frequency range (16 Hz < f < 20,000 Hz). The following are examples of acoustic signals/sounds that Embedded AI can use to solve algorithmic tasks.
More precise selection (optional):
Ultrasonic
Ultrasonic sounds are acoustic signals that lie above the human audible frequency range (20 kHz < f < 1.6 GHz). Below are three examples of ultrasound topics to choose from. If necessary, add your ultrasonic topic to be processed in the input field.
More precise selection (optional):
Optical
Images are created from light reflections or emissions from events, objects or situations in the humanly visible light wavelength range. The following are examples of three image scenarios to choose from. If necessary, add your image scenario to be processed in the input field.
More precise selection (optional):
Infrared
Infrared light is below the range of light wavelengths visible to humans. Below are three examples of infrared scenarios to choose from. If necessary, add your topic to be processed in the input field.
More precise selection (optional):
Wireless
Radio signals are transmitted and received by antennas in a variety of applications. Signals are usually modulated onto radio waves.
More precise selection (optional):
Acceleration
Motion events or the change in motion can be measured by the force of acceleration of a mass/body. The following are examples of three motion scenarios to choose from. If necessary, add your topic to be processed in the input field.
More precise selection (optional):
Forces
Forces and their changes indicate states in technical systems or in vehicles etc. and can be used to control or regulate corresponding processes. Describe your topic.
More precise selection (optional):
Pressure
Pressure can be measured in liquids or gases, from which, for example, conclusions can be drawn about the geographical height of an object or the behavior of technical systems. Describe your topic.
More precise selection (optional):
Sensor
A sensor converts the corresponding physical signal into an electrical signal and then into a digital signal, which can later be analysed by algorithmic signal processing or signal recognition systems.
Signal processing
Signal processing
More precise selection (optional):
Adaptive filters
An adaptive filter in signal processing is a special analogue filter or digital filter that can change its transfer function and frequency independently during operation.
More precise selection (optional):
Signal analysis
Signal analysis is a mathematical method for breaking down signals (time functions or also functions of location and similar) into partial oscillations.
More precise selection (optional):
Stochastic analysis
Stochastic analysis deals with the generalisation of concepts, statements and models from analysis to stochastic processes, i.e. to functions whose values are random.
More precise selection (optional):
Image analysis
Image analysis involves analysing an image systematically, i.e. logically and in an orderly fashion. This can be a painting, a drawing or a photograph, for example.
More precise selection (optional):
Signal transformation
More precise selection (optional):
- Fourier transform coefficients of the transformation of a signal into weighted complex sine/cosine oscillations
- Discrete cosine transformation, used for data compression, for example in the JPEG data format.
- Create an analytical signal, for example to extract the envelope.
- Similar to the FFT, but not on the basis of sine/cosine oscillations, but rather other construction functions, the so-called ‘wavelets’. This allows a variable time-frequency resolution, unlike the FFT, taking into account the uncertainty principle.
- Coordinate transformations (affine transformations) can be used to efficiently implement rotations and scaling for 2D or 3D image signals, for example. This can also be generalised to more dimensions.
- Principal component analysis (PCA) makes it possible to determine a ‘better’ basis for the input data by means of a singular-value decomposition (SVD), which can then be used to realise a dimensional reduction. PCA often precedes classification problems in order to identify more important features in data sets.
- Independant-component analysis (ICA) uses PCA and subsequent rotation to identify components in a data set that are as stochastically independent as possible. ICA is closely related to the topic of BSS (blind source separation).
- More efficient calculation of rotations and translations compared to traditional matrix multiplications.
- The eigenvalue decomposition of a matrix provides information about basis vectors with high variance and low variance (see also PCA).
AI Algorithms
The core of embedded AI technologies is usually an algorithmic system that performs recognition tasks or generates automated decisions.
Machine learning, AI systems and data science technologies are required to create such a system. Such algorithmic technologies are generated in several steps, optimized and brought to the practically required robustness.
In this section, you can find out about our algorithmic technologies and methods and, if possible, make a selection for your project.
![]()
Data Engineering
Data acquisition
Data acquisition refers to all methods of simultaneous or consecutive measurements and counts, including time stamps for measurable or countable data and groups of related data, where applicable.
More precise selection (optional):
Body management
A corpus manager is a tool for multilingual corpus analysis that enables effective searching in corpora.
More precise selection (optional):
Machine Learning Engineering
Software
Both algorithmic and AI systems as well as the technical systems that accompany them generally require the development and operation of software. Professional software development is therefore an important component of embedded AI. Here you can get an overview of our technologies, methods and tools and put together the right building blocks for your project. This includes programming languages, frameworks, platforms, Dev and DevOps methods and tools.
![]()
Languages and frameworks
Markup languages
More precise selection (optional):
Interpreter languages
Source code of interpreter languages is executed directly by the interpreter
More precise selection (optional):
- Python is a high-level programming language used for non-time-critical applications, in particular modelling of signal and image processing algorithms as well as machine learning methods, file work, proof-of-concept systems, etc.
- MATLAB is used for non-real-time modelling of algorithms, in particular signal and image processing, machine learning, data analysis, etc.
- It enables the modelling and simulation of vehicles and their surroundings, including sensor and control technology.
- These source texts are executed by a shell (e.g. Powershell, CMD, bash).
- JavaScripts have access to the browser window and the HTML document displayed in it. Their most important task is to respond to user input in the document.
- Julia develops numerical and scientific computing with high execution speed.
- Lua is an imperative and extensible scripting language for integrating into programmes to make them easier to develop and maintain.
Compiler languages
Source code of compiler languages is compiled before execution.
More precise selection (optional):
Frameworks
Frameworks provide a programming framework and support the development of applications.
More precise selection (optional):
Interface description
An interface description specifies and documents the system interface.
More precise selection (optional):
Embedded Systems
Do you want your embedded AI to run on an electronic board? Then embedded electronics and a software system need to be developed. This usually includes the selection of a suitable processor or microcontroller for the algorithms and software systems, the development, manufacture and, if necessary, series production of an electronics board, as well as the development of an embedded software system. Product design and the production of a housing are also part of this subject area. In this section, you can configure your embedded system so that your product becomes a success story.
![]()
CPU
The centrepiece of your embedded system is the microprocessor (CPU). If you already have requirements or wishes for the CPU, you can describe these in more detail in the following topics.
Typ
More precise selection (optional):
Word width
More precise selection (optional):
Architecture
More precise selection (optional):
Manufacturer
More precise selection (optional):
- Renesas Electronics is a Japan-based company that manufactures and sells semiconductor components and other products for use in the automotive, healthcare, computer peripherals, connectivity, home appliances and industrial markets.
- NXP is the largest semiconductor manufacturer in Europe.
- STMicroelectronics is a leader in a wide range of semiconductor products, including analogue chips, discrete power semiconductors, microcontrollers and sensors.
- Texas Instruments primarily manufactures semiconductors consisting of a wide variety of analogue devices such as power management, operational amplifiers, data converters and digital devices such as digital signal processors, microcontrollers, multi-core processors for industrial and automotive applications.
- Analog Devices is a world-leading analogue and mixed signal technology company that bridges the gap between the analogue and digital worlds with unprecedented solutions.
- Nordic Semiconductors is a Norwegian fabless semiconductor manufacturer. Founded in Trondheim in 1983, the company develops, among other things, chipsets for short-range communication in the ISM band, for example for data transmission in wireless wireless keyboards and mice.
- Infineon is a listed company and the largest semiconductor manufacturer in Germany and one of the ten largest in the world. Infineon offers semiconductor and system solutions focussing on energy efficiency, mobility and security.
CPU performance
‘The ‘Central Processing Unit’ performs the main work in a computer, as it is responsible for processing data and passing on instructions. Accordingly, the speed at which all programmes and processes run depends largely on the performance of the CPU.
More precise selection (optional):
RAM
RAM (Random Access Memory) is your computer's short-term memory, in which the data that the processor is currently using is stored. Your computer can access RAM much faster than data stored on a hard drive, SSD or other permanent storage media.
More precise selection (optional):
ROM/FLASH
More precise selection (optional):
ADC/DAC
ADC and DAC components act as a kind of translator. An ADC has the task of converting an analogue voltage into a binary code at a specific point in time. This means that the ADC samples the analogue voltage and then determines which binary value results at the output of the ADC.
More precise selection (optional):
Interfaces
An interface connects systems that have different physical, electrical and mechanical properties. The definition or specification of an interface contains common properties. This can also include a protocol for communication and data exchange.
More precise selection (optional):
Electronics design
Embedded AI develops your electronics system. The following topics will help you to describe your electronics project in more detail.
Embedded software development
Case
In most cases, you will also need an enclosure for an electronic board. Embedded AI will be happy to develop a housing for your project as follows:
Optics
Optical sensors are used in many applications (e.g. embedded camera sensors). In addition to electronics and software, the development and manufacture of an optical front end (consisting of lens systems and optical mounts) is a prerequisite for the development of an optical sensor system. Embedded AI offers requirements definition, design and testing of optics as well as the optical design and production of lenses and lens mounts with its specialised optics partners.
Connectivity
Connectivity is a crucial aspect for technical systems, as they generally do not work in isolation but need to communicate with other systems and units. This is particularly relevant in areas such as client-server systems, cloud computing, Internet of Things (IoT), mobile applications, applications for smart cities, smart countries or smart regions, sensor networks and Car2X communication. This section lists some technologies and protocols that may be of interest for connectivity in your project.
![]()
Application area
Identifying and structuring the application area of connectivity is important for the use of suitable technologies for communication between technical systems and their optimisation. Do any of the following areas of application apply? If necessary, please add under Other.
Connection
Have you already decided whether the connections in your project will be wireless or wired? If yes, please provide further details. If another option is required, please add the information.
Protocols
Which protocols are required for the project? Please provide details if these have already been determined. Add if other options are being considered.
Protocols
Protocols
More precise selection (optional):
- TCP/IP (Transmission Control Protocol/Internet Protocol) is a basic communication protocol that forms the basis of the Internet and most network communication.
- UDP (User Datagram Protocol) is a connectionless network protocol that enables fast and efficient transmission of data, but without mechanisms for reliability or recovery of lost packets. It is often used in applications that require real-time communication, such as audio/video streaming, online gaming and VoIP.
- gRPC is a modern, cross-platform open source RPC (Remote Procedure Call) framework developed by Google. It uses HTTP/2 as a transport protocol and enables efficient, bidirectional and strongly typed communication between client and server across different languages, supported by protobuf (protocol buffers) for the serialisation of data.
- REST (Representational State Transfer) is an architectural style for communication between client and server via the HTTP protocol that relies on resources, statelessness and the use of standardised HTTP methods such as GET and POST. It promotes simple, scalable and interoperable development of web services and APIs.
- Sockets and WebSockets are mechanisms for communication between client and server via the Internet. Sockets enable bidirectional communication between client and server via network protocols such as TCP or UDP, while WebSockets are a special technology that enable bidirectional, full-duplex communication via an HTTP-like connection, often used for real-time web applications.
- HTTP (Hypertext Transfer Protocol) and HTTPS (Hypertext Transfer Protocol Secure) are protocols that are used to transfer data over the Internet. HTTP defines the rules and methods of how a web client (e.g. a web browser) communicates with a server in order to transfer and display requested data such as web pages or files. HTTPS is a secure variant of HTTP that is protected by encryption (SSL/TLS) to ensure the confidentiality and integrity of the transmitted data.
- QUIC (Quick UDP Internet Connections) is a modern network protocol developed by Google that works on the basis of UDP (User Datagram Protocol). It was developed to improve the performance of HTTP/TCP connections, especially in terms of speed, reliability and security. QUIC offers lower latency by combining connection and data transmission layers that enable multiplexing, error correction and improved handling of network states.
- JSON (JavaScript Object Notation) is a lightweight data format that is used to exchange structured data between a sender and a receiver. It is easy to read and write for humans and machine-readable. JSON is based on key-value pairs and supports various data types such as text, numbers, arrays and objects, making it ideal for communication between applications and web services.
- MQTT (Message Queuing Telemetry Transport) is a lightweight, open protocol for machine-to-machine (M2M) communication that is based on the publish/subscribe model and is optimised for IoT applications.
- MQTT-SN (MQTT for Sensor Networks) is a variant of the MQTT protocol that is optimised for wireless networks with limited resources and sensor devices. It offers similar features to MQTT, but is designed to fulfil specific requirements of sensor networks, including support for low data rates and low power consumption options.
- HDLC (High-Level Data Link Control) is a protocol for data transmission in remote data transmission systems (RDT) and communication networks. It defines a series of rules and procedures for controlling data transmission and error detection. HDLC is particularly known for its reliability and is used in various applications such as point-to-point connections, wireless networks and communication protocols such as SDLC (Synchronous Data Link Control).