Workshop notes
=========

## session 1 Q&A

* RefCal: Regression-based multi-Frequencies Calibration for Crystal-Free micro-Chip (**Yiming Yuan**)
* An Open-Source Multi-Platform SDK for Programming SCuM (**ilip Maksimovic**)
* **Thomas**: why not have UART do through DK?
* **Thomas**: debugging?
* Angle of Arrival Localization for Single-Chip Micro Mote (**Manjiang Cao**)
* Title: SC-Canopy: A Tiered Network Architecture for Indoor Robot Swarms (**Mengyao Liu**)

## Session 2 Q&A

* Ambient BLE (**Jacob Louie**)
* Remind us why isn't the calibration method as fast/robust as Yiming's work
* BLE "geiger counter"?
* What is the approx count per sec you observed
* RIOT port (**Filip Maksimovic**)
* Observatory
* **Thomas** yepkit
* A Supporting Network for Crystal-Free Mesh (**Mitch Montee**)
* **Thomas**: context about the project and water use case
* **Thomas** any feedback about XIAO nRF52840
* Ring Oscillator Frequency Stability Improvement Techniques for Crystal-Free Communication Systems (**David Burnett**)
* poipoi

## Discussion points for session 3

* Nature npj special issue https://www.nature.com/collections/adaadbchbd
* "WireLESS Networking: crystal-free, battery-less and ultra-low power wireless communication"
* T. Chang's AoA tracking paper
* "Internet of micro-Things"
* AoA tracking after introducing SCuM
* interesting CFPs/venues
* ANTS 2026 https://ants2026.org/
* IEEE IFCS 2026 https://2026.ieee-ifcs.org/
* coordination with Kris' team
* Joint project and proposal opportunities
* Review of crystal-free advantages and research goals
* Review of 2023 workshop paper ideas
* Scheduling monthly SCuM Meetings
* Maybe need two interdigitating monthly meetings, held every two weeks, to accommodate all time zones
* This tool can help with planning https://www.timeanddate.com/worldclock/meetingtime.html?month=9&day=22&year=2025&p2=198&p3=195&p4=1232&p1=224&iv=0
* Zoom host w/o waiting room? VU is restricted
* poipoi

## Crystal-free advantages

* No flash needed for security
* No crystal needed for wireless communication
* Re-design the fundamental components in new processes
* High temperature, eg silicon carbide
* Cryogenic applications are "easy"
* Flexible transistor processes
* Implantable/biocompatible processes
* Sensitivity to reciprocal mixing is overblown (paper pending)
* Higher yield through manufacturing one object vs. risk of assembly of PCBs or chiplets
* Fully sealed/passivated package; no environmental contamination
* Oscillator fast start compared to crystal ring-up and lock
* Very wide oscillator tuning range, or pack many oscillators on die easily
*

# Selected Project/Paper Ideas from 2023

(10 of 19 total)

1. DB and Sara: DB intends to measure side-channel attack likelihood using SCuM's CPU 

2. Thomas, Sara, Kris: evaluate algorithms to turn "random" contents of RAM at boot into a string of bytes which remains the same at each reboot

3. DB, Jacob, and Daniel: sharing SCuM-V 23's BLE digital baseband to demodulate and recognize BLE packets after Jacob has exported them digitally via GPIO

5. Tengfei,  Kris get SCuM calibrated through listening for beacon only

6. Tengfei,  Said, can the scum optical receiver can detect lighthouse sweep beam, then calibrate the LC oscillator? Note: lighthouse v1 can be detected.

9. Daniel and Said - lighthouse v2 accelerator design for the Tapeout class.

11. Everyone and tapeout class next year: ideas for new SCuM functions

12. Position paper reviewing crystal-free calibration tecniques and proposing a future standard.    

14. Fil, Titan: sub-us clock synchronization ("time transfer") between two separate SCuM chips?

19. Fil, Titan, Daniel, NFC boot for SCuM-V24

# New paper ideas for 2025+

1. **Thomas**/**Narmin** would love to evaluate https://meshtastic.org/, writing up a "understand the limits of" paper and apply the good aspects of it to SCuM
* PyMesh

1. **Thomas** would love to see mobile lighthouses locate a group of robots
* Heterogeneous robotic and wireless platforms

1. **Thomas** would love to see a (much?) smaller version of the DotBot

1. **Thomas** would like to understand what people could use a 1,000 robot testbed for, this could be a survey of use cases for robots

1. **Dag** would like to have a "swarm robotic hardware as a service", people could reserve a part of the testbed, business model could be based on experiment-minute. "AWS for swarm robotics?" Companies could run real world "simulations" before making a big capital investment decison for example. Allows community to contribute without needing to buy physical hardware.

1. **Narmin** would love to have a version of the testbed that's destined for exploration of unknown locations (or planets).

1. **Narmin** would love to see the DotBots use Swarm Intelligence.

1. **Narmin** would love to see the DotBots have more secure communication.

1. **Fil** wants: a comprehensive build system for replicating experiments (OpenMote + SCuM + nRF + Python for data analysis)

1. **Fil** is designing a self-forming SCuM-only network (no crystals besides silicon)

1. **Sara** would like to develop a robust library in C language for extracting a secret key from SRAM on SCuM using the NVM-free TMVS (Threshold-based Majority Voting Scheme) solution. Integrate this library with RIOT-based cryptographic applications (e.g., AES-128 encryption) using the extracted key.

1. **Tengfei** wants: an automatic process to make batch of boards and test if they works. (@said thinks this is a pretty cool idea)

1. **Tengfei** wants: more tiny packed scum boards to deploy more applications.

1. **Yiming** want to build a micro-scale robot with SCuM integrated(Driven by external field firstly, moving itself secondly)
* The energy for communication
* The bonding stability
* The antenna design

1. **Geovane** and **Luiz** would like to have a mm-sized version of the testbed running on the SCuM, with tiny tiny robots (or just tiny robots in general)

1. **Geovane** is thinking about: would it be possible to have CAIN do robot-to-robot local communication, while Mari does the infrastructure for mobile robots?

1. **Luiz** and **Geovane** would like to work on the RISC-V based SCuM

1. **Jonathan** would like a _fast_ (low-software RF interface on SCuM (< 1 us)

1. **Dag** wonders how feasible it is to do distributed inference with ML on the edge? With video analysis of environment for example?

1. **Mitch** Why can't SCuM hear the nRF52840 openWSN packets?

1. **Mengyao** Minimizing the size of the CaIN tag to jellbean scale for embedding into a conductive silicone-based soft robot that : energy transfer, communication, and sensing.

1. **Jacob** Energy harvesting water buoyes utilizing LoRa mesh (OpenWSN)

1. **Manjiang** Wants to build a infrastructure-less/free wireless based localization system for swarm robots. By communication with each other, robots can locate each other without other infrastructure.