ns-3: Use consistent standards/rates for examples and tests that use Wi-Fi

The change to consider mandatory rates for control responses if a suitable basic rate could not be found (available here for review as Issue 855042) broke the tests involving the wireless/wifi-simple-adhoc-grid example, and the routing-olsr-regression testsuite.

In these cases it is actually the tests that are at fault - in the wifi-simple-adhoc-grid test and the bug780-test part of the routing-olsr-regression testsuite, 802.11b rates were being used with the Wi-Fi PHY standard left at the default of 802.11a. The fix here is to set the standard to 802.11b in the test.

In the tx-regression-test part of routing-olsr-regression, the problem was that the reference traces needed to be updated given the control response rate fix. In this case I have also modified the test to set the standard to 802.11a for clarity.