Como projetar circuitos para sistemas embarcados com poder de processamento limitado?

Um dos pontos mais importantes no momento de escolher o seu microcontrolador é o planejamento do seu projeto. Você precisa ter muito bem definido quais os periféricos que você utilizará. Também será necessário saber o quão complexa será a aplicação que você irá desenvolver.

By focusing on memory optimization, we can significantly enhance the performance of our embedded system. This involves refining the memory subsystem to minimize power consumption, thereby making the system more efficient and effective.

Understand the built-in hardware capabilities of the device. Understand how the peripheral busses and DMA capabilities can come into play. The STM32 allows timers to be linked together to perform quadrature reading quadrature encoders. The ESP32 allows it through the PCNT peripheral. The ESP32 manages to leverage its high clock rate and built-in software to reduce the manufacturing cost and still provide functionality offered in pricier hardware-only solutions. I view the ESP32 as essentially a fancy DMA controller that allows for complex buffer arrangements of 4kb linked lists instead of general circular (ring) buffers. Prefer fixed point math over floating point and avoid unnecessarily using variable sizes that require long arithmetic.

Weigh and balance the project requirements with the MCU and H/W. Could more MCU's balance out the S/W load better and yield simpler code which is more compartmentalized and easier to debug/maintain? For rapid development, use/modify libraries if available, optimize later on. Or if the parameters are tight, start it off by hand, have a better grasp of things. There has to be somebody who is aware of both the coding and the circuitry, S/W & H/W must be in total symphony. Create and re-use libraries for every project. Write lean code. Calculate all possible currents for the circuitry, we don't want the MCU to brownout!

When your processing power is limited, you may want to consider smarter external peripherals or sensors that can monitor or pre-process some of the data on their own. For example, use a temperature sensor with an internal limit check to trigger an interrupt so that your processor does not need to waste cycles polling the status.

A escolha dos componentes externos que você irá utilizar também é parte muito importante ao projetar um sistema com recursos limitados, é nesse momento que você tem que balancear o custo (energético, processamento...) com a confiabilidade do componente, isso dependerá muito do nível de complexidade do seu projeto.

Outro aspecto muito importante é a otimização do código, o desenvolvimento de um código para sistemas embarcados deve ser muito focado em otimização. Evite utilizar bibliotecas desconhecidas e em casos onde se faça necessário utilizá-las, estude a biblioteca e tente abstrair o máximo de funções redundantes e inúteis, sempre que possível.

- Best practice is to do power budgeting while selecting the components. - Configure MCU, sensors, etc to power down mode for low power consumption during inactive state. - Avoid component (active & passive) to be used in the circuit which are prone to leakage current or losses, use better accuracy components. - Use voltage regulators (switching or LDOs) with maximum efficiency & minimum losses to deliver required power.

Design the power supply circuit to be efficient and tailored to the specific requirements of the embedded system. Use voltage regulators or DC-DC converters with high efficiency ratings to minimize power loss and maximize battery life.

Designing circuits for embedded systems with limited processing power involves optimizing algorithms, minimizing resource-intensive operations, and prioritizing power-efficient components. Use low-power microcontrollers and select components with energy-efficient characteristics. Implement efficient coding practices, employ hardware accelerators, and utilize interrupts for task prioritization. Opt for lightweight communication protocols and minimize data transmission. Consider hardware-based processing for computationally intensive tasks. Regularly monitor and optimize code for efficiency. Balancing functionality and resource constraints ensures optimal performance in embedded systems with limited processing power.

You need to also choose the communication type and peripherals carefully considering the speed, bandwidth, physical and software-related constraints.