Data centers, high-tech and aerospace companies, medical suppliers, automotive industry



Operating system for general-purpose AI - in development, beta release expected in 2020


a. hardwired, written exclusively in VHDL for FPGA SiPs containing HBM2 memory, e.g. Xilinx XCVU37P
b. massive parallel processing, for executing exclusively self-programmable dedicated Tipalo neural nets
c. self-programmable dedicated networks on chip - connecting all neural nets including the i/o neural drivers
d. real-time, enabling human-alike reaction times, e.g. 1ms for all simultaneously active neurons
e. made for embedded systems, requires a connectivity to sensors and actors of a corresponding body


a. Interfaces, for Ethernet only, as 10/100/1000 Mb and 10/25/100 Gbit
b. support for HBM2, to be used for storing kernel applications
d. support for DDR4 , to be used for storing accumulated knowledge
d. applications, as brain areas/regions with multiple neural nets connected to an ANS
e. support for self-learning mechanism, which is implemented as the interaction between neural nets

ANS support

a. upload/download, via Ethernet + only for development purposes
b. configuration , as place + route + connect + activate all neural nets within the ANS
c. debugger, as debug, trace and log neural net activities via Ethernet + only for development purposes
d. support for modes, execution pause for certain period of time, self-learning mechanism switch on/off, etc.
e. support for encrypted storage, means the entire ANS is stored internally and encrypted

ANS - Artificial Nervous System, -in development, first prototype expected in 2020

1. requires operating system
for general-purpose AI, details see above
2. neural drivers, for connected sensors, actors and internal organs of an embedded system
3. neural nets
, as applications, which simulate various brain regions
4. self-learning mechanism
, integrated within the neural net application framework
5. accumulated knowledge, with storage and retrieval of information according to situation

ANS software is delivered only within the body hardware of the corresponding embedded system,
with a structure suited strictly for the framework of connected sensors, actors, internal organs,
where the skills, abilities and knowledge are implemented according to the client requirements.

ANS classification

There are different levels of intelligence, therefore we have the following classification,
where, in terms of capacity, each level is ca. 1000 times greater than the previous level:

Level 1, corresponds to insects, enables the following skills:

              - simple sensor input drivers, allows perception of different elements
                e.g. visual black+white, etc.
              - simple actor output drivers, allows commands of simple actions of body limbs,
                e.g. locomotion with small speed, etc.
              - simple glue logic, allows pre-defined reflexes for processing as input to output connectivity
                e.g. fixation on sensor element with certain attribute(s) and follow the trail accordingly
              - internal representation of entire body, as structure of connected body parts
                e.g. map of all body components with connectivity between them
              - internal connectivity, between body components and sensors + actors
                e.g. sensors + actors + internal organs

Level 2, corresponds to mammals/fishes/birds, enables the following skills:

              - complex sensor input drivers, allows perception of objects
                e.g. color visual as RGB
              - complex actors output drivers, allows commands of complex body actions,
                e.g. locomotion with different speed, sense of balance, coordination of own body parts, etc.
              - memory, means storage of many and different objects and landscapes
                e.g. objects can be grouped, which allows a classification based on user-defined templates
             - complex glue logic, allows multiple pre-defined own status quo induced by situations
                e.g. situation triggers imminent danger caused by a crowded landscape
              - self-learning mechanism, allows user-defined behaviour based on multiple factors
              - e.g. analysis of a situation with different objects within a certain landscape

Level 3, corresponds to humans, enables the following skills:

                 - language processing, assuming 2 different pre-defined specific applications
                    e..g. center for speech interpretation, center for speech production
                 - accumulation of knowledge, by using language processing
                    e.g. physical, logical and abstract terms for building and using an user-defined dictionary
                 - imagination, requires retrieval on demand of memorized objects and landscapes
                    e.g. internal representation and processing of possible scenarios
                 - complex logic processing, which includes selection and association of terms and actions
                    e.g. physical, logical and abstract terms for building and using an user-defined encyclopedia
                 - self-awareness, recognition of own body within the internal representation of perceived objects
                    e.g. proved via the mirror self recognition test


1. Some animal species are self-aware, means recognizing themselves in a mirror
          -  e.g. mammals like elephant, bonobo, birds like Eurasian magpie, fishes like cleaner wrasses, etc.
2. Self-aware does NOT mean free will
3. Free will does NOT imply one has an inconsiderate and aggressive behaviour towards the entire world

Templates for product lines

For each of the 3 levels of intelligence, there are different digital templates  for each ANS:

Manager of static facilities, like intelligent buildings, automated factories, etc.
      - implemented as a static group of connected Level 1 embedded objects

Pilots for vehicles of all types, for trucks, ships, planes and satellites
      - implemented as different Level 2 embedded objects, according to each type of locomotion

Robotic workers in harsh environments, as robots for activities in space or other complex environments
      - implemented as Level 3 with different levels of knowledge, e.g. human of different age and specialty

Program for Early Adopters / Custom designs

Early Adopters and custom designs can be used to develop specific product lines, based on the templates.

For each inquiry there will be a project, with the following steps:

1st step is feasibility study, will include the requirements for the desired:

a. body hardware, with corresponding sensors/actors/internal organs
b. environment, to be used in, e.g. terrestrial, naval, aerial, space, etc.
c. skills, needed to perform certain activities
d. knowledge, accumulated via specific training

2nd step is prototype creation, with all sensors/actors/internal organs and body framework
3rd step is specific training, to obtain the skills and knowledge
4th step is test, possible under real conditions