Difference between revisions of "Ultrasound Low power WiFi with IMX7"
From iis-projects
(→Status: Available) |
|||
Line 16: | Line 16: | ||
* Experience with Heterogeneous platforms is nice to have | * Experience with Heterogeneous platforms is nice to have | ||
− | ===Status: | + | ===Status: Completed === |
− | + | Guillaume Thivolet | |
: Supervision: [[:User:Vsergei|Sergei Vostrikov]], [[:User:Cosandre|Andrea Cossettini]] | : Supervision: [[:User:Vsergei|Sergei Vostrikov]], [[:User:Cosandre|Andrea Cossettini]] | ||
Revision as of 13:56, 11 September 2021
Contents
Short Description
In a wireless ultrasound probe, transmitting raw ultrasound data at high frame rates with low power consumption appears as a challenging task. On the one hand, WiFi solutions compatible with low-power embedded systems (e.g. MCUs) are typically limited to small throughputs (< 5 Mbps). On the other hand, faster WiFi modules that are optimized for high-performing SoCs (running a general-purpose OS) come with increased power consumptions. A promising solution is to use heterogeneous multicore platforms, combining the abvantages of MCUs and high-performance SoCs, and featuring efficient power-down modes.
Goal & Tasks
The goal of the mini-project is to explore a processor i.MX 7ULP from NXP) for low-power ultrasound data streaming through WiFi. In this mini-project you will:
- Program real-time domain (Cortex-M4, FreeRTOS) of IMX SoC to receive raw ultrasound data through QSPI + DMA
- Program application domain (Cortex-A7, Linux) to stream the data through WiFi.
- Test the throughput of the system and corresponding power consumption.
Prerequisites
- Python (sklearn, tensorflow)
- C
- Experience with Heterogeneous platforms is nice to have
Status: Completed
Guillaume Thivolet
- Supervision: Sergei Vostrikov, Andrea Cossettini
Character
- 20% Literature Study
- 80% Coding