Conclusions and Future Work
This chapter sums up the outcomes of the project and gives directions for future work.
Overall the project has been successful.
‘Black Spot’ camera module was designed
and three modules were constructed.
n initial firmware implementation was completed
showing that each module can track 27 markers at 60 fps. Further firmware development is required. Of particular note is the implementation of the marker quality algorithm. This algorithm could allow the hub to give a lower weighting to markers that have poor cen-
troid standard deviations.
lso, due to changes in the internal operation of the firmware,
the partitioning algorithm must be re-implemented.
The hardware topology and choice of Blackfin DSP and
ptina CMOS image sensor has
proven to work well. However, the initial design required a number of hardware modifi- cations that should be integrated into any new designs. The Blackfin DSP has proven to be fast enough to allow the tracking algorithms to run in real time. However, in the current implementation, the bandwidth of the communications bus is a limiting factor. The num- ber of centroid packets that can be sent is approximately 28 per frame. This limitation is shared between the three camera modules.
The decision not to use SDR M has allowed functioning modules to be produced that were not a certainty in a design using SDR M. However, the lack of SDR M has made firmware development more difficult. Future revisions of the PCB could use SDR M to allow the firmware to be tested easily.
The PCB’s dimensions could be reduced in future revisions by decreasing the sizes of the
debug and communications headers and by optimising part placement.
Blackfin BF533 DSP is available in a smaller ‘chip scale’ BG
dditionally, the Furthermore, a