Independent groups is one type of experimental design, whereby different participants take part in each condition of the independet variable. Random allocation is used when possible to ensure that there is no systematic difference between the participants' characteristics in each group.
Repeated measures is another type of experimental design. This is where participants take part in every condition of the independent variable and are acting as their own controls. To make sure the results are not affected by order effects (e.g. boredom or fatigue), we can counterbalance the conditions so, for example, one group of participants take part in condition A then B, and another group takes part in B then A.
Matched pairs combines the advantages of both an independent groups design with a repeated measures design. Each participant is matched as closely as possible with a participant in the other condition, so in effect researchers can treat their results as if they were the same person but without the worry of order effects.