Meredith Beveridge Lecocke

1
Smart Sensors

• Meredith Beveridge Lecocke
• Schlumberger

2
Smart Sensor Concept
Differences
Constraints

• About 18 smart sensors running identical code
  slight differences indicated by downloaded
  configuration

Components
Processes
Drivers
Diff Filter Buffer
Tight timing
microcontroller
Small RAM ROM
Sensor
ADC
SPI
Non-volatile Memory
Interpolate
Filter
ADC
Sensor
config
status
buffer
firmware update
Timers
Diff ADCs
Watchdog
TDMA bus
Debug port
RS485
RS485
Diff TDMA slots
Tight timing
3
Model Driven Development for Embedded Systems

• Vassilis Varveropoulos
• Senior Software Engineer
• Schlumberger

4
Model Driven Development

• MDD is becoming an attractive alternative as
  embedded systems became more complex.
• Extra level of abstraction by developing system
  using UML model only.
• Ideal use of MDD is to develop only through the
  model, code will be an artifact of the compiler.
• Many commercial tools in the market, but not
  fully mature and flexible enough yet.
• Model compilers are still in their infancy.

5
MDD Research Proposal, Phase 1

• Evaluate existing embedded MDD tools (e.g.
  Rhapsody, RoseRT, Artisan RTS, TauG2).
• As a baseline, two representative systems from
  SLB could be used (one controller and one DSP?).
• Cut-down version of the system software could be
  developed using MDD.
• Identify limitations and shortcomings of
  commercial tools.
• Findings will define phase 2 (if one is
  necessary).

6
MDD Research Proposal, Phase 2

• Based on the findings of phase 1, propose and
  develop possible solutions.
• We should avoid creating a new MDD tool if
  possible.
• We might find that simply developing a better and
  more flexible model compiler is what is required.
• A commercial UML tool could be used to develop
  the model.
• Phase 1 will require a 3 month internship, Phase
  2 will be at least a 1-2 year project.

7
Motor Control Electronics

• Sridhar Sana
• Schlumberger

8
Software Architecture

• Code written in C for ADSP2181
• Fixed point 1.15 arithmetic used for control
  algorithm computation
• Resolver-to-digital conversion
• Park and Clark transformations
• Position PID, Torque/Current PIDs
• Space vector modulation
• Control system sample rate is 139uS (hard
  real-time constraint)
• RS-485 Communication(soft real-time constraint)

9
Expectations from internship

• Explore RAP to optimize present code to meet
  control system real-time constraint
• Come-up with design process/best-practices for
  writing optimal C code for embedded DSPs