What is What and Why



Machine Learning

What?

Machine learning is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience (training data) without being explicitly programmed. Machine learning focuses on the development of computer programs that can access data and use it to learn for themselves.

Mathematically,

Learning a target function (f) that best maps input variables (X) to an output variable (Y). Y = f(X) This is a general learning task where we would like to make predictions in the future (Y) given new examples of input variables (X).

Why?

  • It can be used for the classification of disease in the plant/human body, predicting the wear/tear of the machine, predicting the failure, optimizing power and area on chip based on routing recommendation, and so on. 

Internet of Things (IoT)

What?

Internet means any network wired/wireless (e.g. Wi-Fi, Bluetooth, access points).

Things mean any electronic device with a built-in sensor.

So basically any electronics device with a sensor built-in connection to the internet is IoT.

Why?

  • It can be used to get real-time data from the field about moisture, humidity, NPK content.
  • It can be used to provide real-time health information with wearable sensors
  • It can be used to provide real-time distance measurements and feedback for vehicles and so on.

Embedded Systems

What?

Embedded is simply a piece of code.

The system can be an electronics device with memory, timer, I/O, clock, processor/FPGA.

So basically embedding a piece of code in a system is an embedded system.

Why?

  • It can be used to perform a real-time action/event without failing (For e.g. clicking a picture by pressing a button, printing/scanning on pressing a button, setting up a temperature in micro-ovens, moving robots to a precise position, and so on)

Edge Computing

What?

Edge means a node that is sitting in the proximity of any machine/device and collecting the data.

Computing means processing the data and proving feedback rather than sending the data to the cloud.

So, in Edge computing data is processed by the device itself or by a local computer or server, rather than sending to a cloud.

Why?

  • Suppose you want to predict the wear/tear of machines. You simply place the edge device in the proximity which will collect the data and processes to give feedback on maintenance and failure. 

Digital Twin

What?

Twin means replicas of any physical device, process, product.

Digital means using sensor data from embedded machines to create accurate, real-time simulations.

So basically creating a virtual model of any physical device, process, or product for various purposes is Digital Twin.

Why?

  • It allows data scientists and engineers to run simulations before actual devices are built and deployed.
  • It allows assessment with actionable insights into machine performance and maintenance needs.

Computer Vision

What?

Vision refers to any digital camera/image sensor used for acquiring image/video.

Computer refers to the computational device with artificial intelligence that trains to interpret and understand the visual world.

So basically it deals with how computers can gain a high-level understanding of digital images or videos

Why?

  • It allows computers to understand and label images, driverless car testing, daily medical diagnostics, and monitoring the health of crops and livestock.

Sensors

What?

An electrical, optoelectrical, or electronic device that interacts with the environment to sense the light, temperature, pressure, moisture, distance, vibrations, etc. The frontend is the analog electronics that sense these parameters and provides the electronics signals. Electronic sensors can detect everything from light to distance to acceleration. 

Why?

  • It can be used to detect events or changes in its environment and send the information to other electronics of frequently computer processors

FPGA 

What?

FPGA: It has a sea of Gates that can be re-programmed. Suppose you have programmed the FPGA to add two numbers and deployed it in the field and now the requirement is to add 4 numbers, so basically you can re-program the logic gates in the field to meet the requirements.

FPGA is complex and widely used across industries because of its scalability, performance, low power.

Simply understand FPGA as a series of logic blocks that can be programmed for any application.

Why?

  • It provides higher performance, lower power, and scalability.
  • It can be used for complex applications, even you can build 32-bit processors on FPGA.
  • It is easy to deploy in the field and no tapering or fabrication required.





Comments

Post a Comment

Popular posts from this blog

FPGA In Layman''S Terms

Image
FPGA is an Integrated Circuit which can be reconfigured again and again to perform different tasks as desired. For example, an FPGA can be configured to function as a processor, and then it can be reconfigured to function as a Graphics Processing Unit (GPU), or hardware video encoder, or something else. The point is that FPGA can theoretically be programmed to function as any other Digital Integrated Circuit. Think of FPGA as a breadboard for Digital Circuits. All the gates and flip-flops are present on the FPGA, and there are wires going through the complete chip. Circuits are made by connecting these wires to the relevant gates or flip-flops as per intended design. Instead of connecting the gates/flip-flops using a physical wire as in a breadboard, FPGAs has programmable interconnects that can be rewired programmatically. There are dedicated wires/routes for clock signals and only a selected number of FPGA pins are allowed to drive those global clock routing wires. FPGA Vs Microcontr...
Read more

BOARD BRING-UP

Board bring-up is the process of checking and validating a PCB assembly for the first time. The aim is to make sure that the board has been manufactured and assembled correctly and that all of the subsections of the design work as intended. This can be a nervous time for an electronics designer as it is a key proof point that many weeks or months of work has been successful. Note that bring-up is not the same as validating an entire product or system – it is usually limited to validating the hardware. Writing firmware will likely be a part of the process, but only to the extent required to ensure the processor, its peripherals and external circuitry connected to them are working as expected. Board Bring-up process. 1. Visual Inspection: Print out assembly drawing with all the components and test points to have a quick reference during validation. All the components are placed correctly. All the PINs 1 are placed correctly. There are no visible short circuits between PADS. Are pads sold...
Read more

High Speed Digital Design Considerations

Image
The four main consideration of today’s high-speed design:                                                                                                                                                            Ref(1) We need to consider signal integrity , which includes parameters such as Overshoot and Ringing. Need to look at crosstalk which is about EMI coupling between traces on PCB board Timing is about calculating the flight time EMI and radiation  These parameters are related to each other and aff...
Read more