I've used the MGCV package to fit a constrained regression where the coefficient could not be negative: 

[Constrained Regression][1]

Also the 'quadprog' and solve.QP maybe useful.

Both, however, have a little bit of a learning curve, at least for me.


  [1]: http://www.stat.columbia.edu/~tzheng/tianblog/2006/07/fitting-constrained-least-square.html