It can't actually be done, except in the most schematic way which discards the structure.
In the first case, the three sentences can be represented as follows, using the predicate $Fool(x,y,t)$ to mean "$x$ can fool $y$ at time $t$":
$$\begin{align}
(\forall x)(\exists y)(\forall t)\,Fool(x,y,t) \tag{1a} \\
(\forall x)(\forall y)(\exists t)\,Fool(x,y,t) \tag{1b} \\
\neg(\exists x)(\forall y)(\forall t)\,Fool(x,y,t) \tag{1c}
\end{align}$$
Note that (1c) is equivalent to:
$$
(\forall x)(\exists y)\neg(\forall t)\, Fool(x,y,t) \tag{1c'}
$$
All of these fine distinctions go away if you represent the sentences in propositional logic. The sentences have no logical connectives, so they must be represented as simple propositional variables $p,q,r$ (or $\neg r$, if you prefer — it really doesn't matter) respectively. Notice that (1a) and (1c) together imply that (if there are any people at all, $x$, then) there are at least 2 people ($y$) :) But that can't be deduced from $p,q,r$, and can't even be expressed in propositional logic.
2) "Peter has at least 2 children". If $C(x,y)$ is a predicate meaning that "$x$ has a child $y$", and if $\mathsf{P}$ is a constant denoting Peter, then in first order logic the sentence can be represented by:
$$
(\exists x)(\exists y)\,(C(\mathsf{P},x)\land C(\mathsf{P},y)\land x\ne y) \tag{2}
$$
In propositional logic, this has to be represented by a single propositional variable $p$: the sentence has no truth-functional connectives, no propositional structure. From (2), we can deduce that Peter has children (at least one child), and that Peter does not have only one child; but $p$ is a black box from which nothing follows except $p$.
It's a bit of a trick question, and as Ove points out in his answer, perhaps its intent is to show the limitations of propositional logic without variables and quantifiers for individuals.